Rapid prediction method for temperature field and stress field of dangerous part of key component of small turbojet engine

Abstract

The invention relates to a rapid prediction method for a temperature field and a stress field of a dangerous part of a key component of a small turbojet engine, belongs to the field of optimization design of small turbojet engines, and solves the problems of high cost and low efficiency of simulating the temperature field and the stress field of the turbojet engine by an existing conventional numerical simulation method. Comprising the steps of establishing a turbojet engine key component geometric model; generating temperature field and stress field data sets under different working conditions; reducing the order by using a POD flow field order reduction model and obtaining a transformation relation; training the deep learning model to obtain a mapping relation; inputting to-be-calculated working condition data to obtain reduced-order data; the distribution of the temperature field and the stress field under the specific working condition is obtained through recalculation by means of the transformation relation, rapid prediction can be achieved, and cost is saved. According to the method, deep learning and a flow field reduced order model are adopted, local dangerous parts under the given working condition can be predicted in time, calculation and time cost is saved, and timely data feedback is provided for prolonging the service life of an engine and keeping good performance of the engine.

Description

Technical Field

[0001] The invention relates to a method for quickly predicting temperature fields and stress fields of dangerous parts of key components of a small turbojet engine, and belongs to the field of small turbojet engine optimization design. Background Art

[0002] A small turbojet engine is an engine that uses high-speed jets to generate thrust to propel aircraft. According to the order in which the gas circulates in the engine, its main structural components include the air intake, compressor, combustion chamber, turbine and tail nozzle. As a powerful power device with high energy density, the small turbojet engine has the advantages of small structure, light weight and sufficient power compared to traditional engines. It can be widely used in the fields of micro-UAVs, aircraft, auxiliary control devices, etc.

[0003] When the engine is working, sometimes in order to meet certain working needs or when the engine is in a relatively harsh working environment, the engine will run at a high speed and high injection volume close to the limit state. At this time, all components of the entire engine structure are subjected to high temperatures and thermal loads. The yield strength and tensile strength of the engine component materials will be greatly reduced due to high temperature conditions, and high-temperature creep will occur under long-term conditions. At this time, the service life and working performance of the turbojet engine will be seriously affected. Even in normal working conditions, due to the uneven distribution of the internal temperature field, local components will have high temperatures. For example, in a turbojet engine, due to the combined influence of the combustion chamber's own structure, combustion organization and cooling mechanism, the temperature field at the outlet of the combustion chamber will form an outlet flow field with obviously uneven and circumferential and radial temperature gradients, and the temperature of the local gas core area will be significantly higher than that of the surrounding fluid. The uneven temperature distribution at the outlet of the combustion chamber will cause the turbine blades to receive uneven heat loads, which will make the life of the local turbine blades much lower than that of other parts. As a relatively complex organic whole, problems with local components of the engine will also affect the life and smooth operation of the entire engine.

[0004] Therefore, in order to extend the life of the engine and ensure the working performance of the engine, it is necessary to quickly predict the temperature field and stress field of the engine under various operating conditions and pay attention to the most dangerous key component parts in the entire engine so as to change or optimize the operating state of the engine in a timely manner. Although the commonly used numerical simulation method can accurately simulate the temperature field and stress field under various operating conditions, the calculation cost and time cost are relatively large, and the rapid prediction of the temperature field and stress field of the engine cannot be achieved. Therefore, the present invention proposes a method for quickly predicting the temperature field and stress field of the dangerous parts of the key components of a small turbojet engine. Summary of the invention

[0005] The present invention proposes a method for quickly predicting the temperature field and stress field of dangerous parts of key components of a small turbojet engine. By integrating and applying advanced monitoring technology, the present invention solves the problem of high cost and low efficiency of simulating the temperature field and stress field of a turbojet engine by existing conventional numerical simulation methods.

[0006] A method for quickly predicting temperature fields and stress fields at dangerous parts of key components of a small turbojet engine, the method comprising the following steps:

[0007] S1. Establish a geometric model for the dangerous parts of a given turbojet engine key component;

[0008] S2, generating different working condition data and corresponding training data sets and test data sets of temperature fields and stress fields, wherein the generated data sets include temperature field data and stress field data;

[0009] S3, using the POD flow field reduction model to reduce the temperature field data and the stress field data, and obtain the transformation relationship between the original flow field data and the reduced-order data;

[0010] S4. Using the working condition data in the training data set and its corresponding temperature field and stress field reduced-order data, the deep learning model is trained to obtain the mapping relationship between the working condition and the temperature field reduced-order data and the stress field reduced-order data respectively;

[0011] S5. Input the working condition data of the temperature field and stress field distribution to be calculated into the trained deep learning model to obtain the reduced-order data of the temperature field and stress field;

[0012] S6. Recalculate the temperature field and stress field reduced-order data using the transformation relationship between the original flow field data and the reduced-order data to obtain the temperature field and stress field distribution under specific working conditions.

[0013] Furthermore, in S1, the key component is a part that is subjected to high temperature, high pressure and complex stress during the operation of the turbojet engine.

[0014] Furthermore, in S2, the operating condition data is selected in a range between a turbojet engine startup state and a limit operation state.

[0015] Furthermore, in S2, among the selected operating condition data, the combined operating condition data includes fuel flow, intake pressure and engine speed.

[0016] Furthermore, in S2, it also includes dividing the geometric model of the key components of the turbojet engine using unstructured grids.

[0017] Furthermore, in S2, the temperature field and stress field distribution of key components of the turbojet engine under different working conditions are numerically simulated to obtain temperature field and stress field data of a series of working condition data.

[0018] Furthermore, in S2, the operating condition data and its corresponding temperature field and stress field distribution data are divided into a training data set and a test data set, wherein the training set accounts for 0.8.

[0019] Furthermore, in S3, the following steps are also included:

[0020] S31, respectively obtaining a temperature field average value and a stress field average value for the temperature field data and the stress field data;

[0021] S32, standardizing the temperature field and stress field data using the obtained temperature field average value and stress field average value;

[0022] S33, calculating the covariance matrix of the standardized temperature field and stress field data;

[0023] S34, calculating the eigenvalues ​​of the covariance matrix and sorting them in descending order;

[0024] S35, taking the eigenvectors corresponding to the top-ranked eigenvalues ​​to form a mapping matrix;

[0025] S36. By multiplying the temperature field and stress field data matrix on the left by the mapping matrix, low-order temperature field and stress field data, i.e., fundamental modal coefficients of the flow field, are obtained, thereby reducing the order of the temperature field data and the stress field data using the POD flow field reduction model, and obtaining the transformation relationship between the original flow field data and the reduced-order data.

[0026] A storage medium stores a computer program, which, when executed by a processor, implements a method for quickly predicting the temperature field and stress field of a dangerous part of a key component of a small turbojet engine.

[0027] A computer device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for rapidly predicting the temperature field and stress field of dangerous parts of key components of a small turbojet engine as claimed in the claim.

[0028] Beneficial effects of the invention: The method for quickly predicting the temperature field and stress field of dangerous parts of key components of a small turbojet engine of the invention adopts a deep learning model and a flow field reduction model, which can realize timely prediction of the temperature field and stress field of local dangerous parts of a turbojet engine under given working conditions, avoiding the problems of high cost and low efficiency caused by the conventional numerical simulation method in the process of simulating the temperature field and stress field, effectively saving computing cost and time cost, and providing timely data feedback to extend the life of the engine and maintain good working performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A flowchart of a method for rapidly predicting temperature fields and stress fields at dangerous locations of key components of a small turbojet engine according to the present invention;

[0030] Figure 2 The figure is a calculation flow chart of flow field reduction, prediction and reconstruction in one embodiment of the present invention. DETAILED DESCRIPTION

[0031] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] Reference Figure 1-Figure 2 As shown, a method for quickly predicting the temperature field and stress field of a critical part of a small turbojet engine comprises the following steps:

[0033] S1. Establish a geometric model for the dangerous parts of a given turbojet engine key component;

[0034] S2, generating different working condition data and corresponding training data sets and test data sets of temperature fields and stress fields, wherein the generated data sets include temperature field data and stress field data;

[0035] S3, using the POD flow field reduction model to reduce the temperature field data and the stress field data, and obtain the transformation relationship between the original flow field data and the reduced-order data;

[0036] S4. Using the working condition data in the training data set and its corresponding temperature field and stress field reduced-order data, the deep learning model is trained to obtain the mapping relationship between the working condition and the temperature field reduced-order data and the stress field reduced-order data respectively;

[0037] S5. Input the working condition data of the temperature field and stress field distribution to be calculated into the trained deep learning model to obtain the reduced-order data of the temperature field and stress field;

[0038] S6. Recalculate the temperature field and stress field reduced-order data using the transformation relationship between the original flow field data and the reduced-order data to obtain the temperature field and stress field distribution under specific working conditions.

[0039] Specifically, first, the present invention generates a data set covering different working conditions (including starting to the extreme operating state, involving multiple parameters such as fuel flow, intake pressure and engine speed, etc.), and combines the POD flow field reduction model to reduce the temperature field and stress field data, which not only reduces the amount of data and computational complexity, but also greatly improves the prediction efficiency and reduces the computational cost under the premise of ensuring that key information is not lost, thus overcoming the problems of high computational cost and low efficiency in the prior art. Secondly, the deep learning model is used to train the training set containing the reduced-order data, which can accurately learn the changing laws of the temperature field and stress field under different working conditions, thereby ensuring the accuracy of the prediction results and providing a reliable basis for the state assessment of key components of the engine. Finally, the method can adapt to a variety of working conditions, quickly and accurately obtain the temperature field and stress field information of the dangerous parts of the key components, help to timely discover potential hidden dangers and failure risks, and provide strong support for engine maintenance, including formulating reasonable maintenance plans, taking preventive measures in advance, etc., thereby extending the service life of the engine, improving reliability and safety, and reducing operating costs.

[0040] Furthermore, in S1, the key component is a part that is subjected to high temperature, high pressure and complex stress during the operation of the turbojet engine.

[0041] Specifically, the components of the turbojet engine are reasonably simplified, and this simplified operation will greatly improve the calculation efficiency. By removing some complex structures and details that have little effect on the temperature field and stress field of key components, the calculation process is made more efficient and unnecessary consumption of computing resources is reduced. The geometric model is established using existing commercial modeling software, which makes full use of the advantages of commercial software. These software usually have powerful modeling functions and wide applicability, and can accurately construct the geometric shapes of the key components of the turbojet engine, providing an accurate basic model for subsequent temperature field and stress field analysis. This not only ensures the quality of the model, but also saves the time and cost of developing modeling tools by yourself. In this embodiment, the basic structure of the small turbojet engine consists of a single-stage compressor and a single-stage turbine blade, and its main key components mainly include a combustion chamber, a turbine blade, a fuel nozzle, etc. The above-mentioned key components are subjected to high temperature, high pressure and complex stress during the operation of the engine. Research on these key parts can more accurately focus on the areas where the engine is most prone to problems, making the prediction and analysis of the temperature field and stress field more targeted and of practical application value, and helping to timely discover the potential safety hazards and failure risks of these key parts, thereby providing strong support for the maintenance and optimization of the engine.

[0042] Furthermore, in S2, the operating condition data is selected in a range between a turbojet engine startup state and a limit operation state.

[0043] Specifically, on the one hand, the range selected in this embodiment comprehensively covers the special operating conditions from startup to extreme operating conditions in the actual operation of the engine, ensuring that the generated data set fully reflects the changes in the temperature field and stress field of key components under different operating conditions, avoiding inaccurate predictions due to incomplete operating conditions, and making the prediction model trained based on this data more versatile and accurate. It can better adapt to the complex and changeable actual operating environment throughout the life cycle of the engine, whether it is structural design and material selection in the design stage, or problem discovery in the operation stage and strategy formulation in the maintenance stage, and provide strong support for the performance and reliability management of the engine. On the other hand, this comprehensive coverage of operating conditions helps to discover potential problems in a timely manner, extend the service life of the engine, and improve overall performance and economy.

[0044] Furthermore, in S2, among the selected operating condition data, the combined operating condition data includes fuel flow, intake pressure and engine speed.

[0045] Specifically, in this embodiment of the present invention, operating condition data must first be selected. In order to cover various operating conditions of the turbojet engine as much as possible, a series of combined operating condition data are selected within the data variation range between the engine start-up state and the extreme operating state. Among them, more combined operating condition data should be selected near the data of the common operating conditions of the engine, including fuel flow, intake pressure, engine speed, etc.

[0046] In this embodiment, the combined working condition data include fuel flow, intake pressure and engine speed, and are selected within the data variation range between the engine start state and the extreme operating state, and more combined working condition data are selected near the common operating state data of the engine. This approach comprehensively covers various working states of the turbojet engine, so that the generated data set can accurately reflect the temperature field and stress field changes of key components under different working conditions. By covering these key parameters, it is possible to have a deeper understanding of the working characteristics of the engine under different fuel flow, intake pressure and engine speed, and provide richer and more accurate information input for the prediction model. In practical applications, it helps engineers to better optimize the structure and select materials in the engine design stage, timely discover potential problems and perform accurate fault diagnosis in the operation stage, and formulate a more scientific and reasonable maintenance plan in the maintenance stage, thereby improving the performance, reliability and economy of the engine and extending its service life.

[0047] Furthermore, in S2, it also includes dividing the geometric model of the key components of the turbojet engine using unstructured grids.

[0048] Specifically, in this embodiment, for the geometric model established in S1, an unstructured grid is selected for grid division and the grid independence is verified, the calculation boundary conditions are determined, and the relevant calculation models in the commercial numerical simulation software are set, wherein the turbulence model adopts the gas phase NS (Navier-Stokes) equation of the standard k-ε model, the combustion model adopts the PDF (Probability Density Function) non-premixed combustion model, and the spray model adopts the DPM (Discrete Phase Model) model. Then, the temperature field and stress field distribution of the key components of the turbojet engine under different working conditions are numerically simulated to obtain the temperature field and stress field data of a series of working condition data.

[0049] The selection of unstructured grids for grid division and independence verification ensures the rationality and accuracy of grid division and improves the accuracy and reliability of numerical simulation. Determining the calculation boundary conditions and setting the relevant calculation models in the commercial numerical simulation software, such as using the gas phase NS equations of the standard k-ε model as the turbulence model, the PDF non-premixed combustion model as the combustion model, and the DPM model as the spray model, can more accurately simulate the complex flow, combustion and spray processes of the turbojet engine under different working conditions. By numerically simulating the temperature field and stress field distribution of key components under different working conditions, a series of temperature field and stress field data of working condition data are obtained, which provides rich and accurate basic information for subsequent analysis and prediction. In practical applications, this helps engineers to have a deep understanding of the temperature and stress changes of key components of turbojet engines under various working conditions, providing a strong basis for the design optimization of the engine, improving the reliability and safety of the engine, and extending its service life.

[0050] Further, in S2, the temperature field and stress field distribution of key components of the turbojet engine under different working conditions are numerically simulated to obtain a series of temperature field and stress field data of working condition data. The working condition data and the corresponding temperature field and stress field distribution data are divided into a training data set and a test data set, wherein the training set accounts for 0.8.

[0051] Specifically, in this embodiment, a series of operations performed on key components of a turbojet engine have achieved remarkable results in terms of computational efficiency savings. Numerical simulations of temperature field and stress field distributions of key components of a turbojet engine under different operating conditions are performed to comprehensively obtain detailed status information of key components under various operating conditions, thereby laying a solid data foundation for subsequent analysis and prediction. By dividing the operating condition data and its corresponding temperature field and stress field distribution data into a training data set and a test data set, and setting the training set ratio to 0.8, a reasonable allocation of data resources is achieved. A large proportion of training sets enables the deep learning model to fully learn the data characteristics and laws under various operating conditions, greatly improving the accuracy and generalization ability of the model, and reducing the time and cost of repeated calculations due to inaccurate models. Independent test data sets can objectively evaluate model performance, ensure the reliability of the model in practical applications, and avoid erroneous decisions and subsequent correction costs caused by unreliable models. This data division method of this embodiment helps to more accurately predict the temperature field and stress field of key components of a turbojet engine under different operating conditions in practical applications, providing a scientific basis for engine design optimization, fault diagnosis, and maintenance decisions, thereby improving engine performance and reliability, and reducing operating costs and risks.

[0052] Furthermore, in S3, the following steps are also included:

[0053] S31, respectively obtaining a temperature field average value and a stress field average value for the temperature field data and the stress field data;

[0054] S32, standardizing the temperature field and stress field data using the obtained temperature field average value and stress field average value;

[0055] S33, calculating the covariance matrix of the standardized temperature field and stress field data;

[0056] S34, calculating the eigenvalues ​​of the covariance matrix and sorting them in descending order;

[0057] S35, taking the eigenvectors corresponding to the top-ranked eigenvalues ​​to form a mapping matrix;

[0058] S36. By multiplying the temperature field and stress field data matrix on the left by the mapping matrix, low-order temperature field and stress field data, i.e., fundamental modal coefficients of the flow field, are obtained, thereby reducing the order of the temperature field data and the stress field data using the POD flow field reduction model, and obtaining the transformation relationship between the original flow field data and the reduced-order data.

[0059] Specifically, in this embodiment, first, by successively obtaining the mean value of the temperature field and stress field data, performing standardization operations, calculating the covariance matrix, sorting the eigenvalues ​​in descending order, and selecting the front eigenvectors to form a mapping matrix, and finally using the mapping matrix to obtain low-order data, the POD flow field reduction model is used to reduce the order of data. This process greatly reduces the amount of data and the complexity of calculation, and avoids the high calculation cost and low efficiency caused by a large amount of data in the traditional method. At the same time, the key flow field information is also retained to ensure that important data features are not lost while reducing the calculation burden. Secondly, this series of steps enables more efficient data processing and significantly improves the calculation speed when analyzing and predicting the temperature field and stress field of the key components of the turbojet engine. Compared with the traditional method, there is no need to spend a lot of time and computing resources to process a huge amount of data, which greatly shortens the time required for prediction. Furthermore, the transformation relationship between the original flow field data and the reduced-order data obtained builds an important bridge for subsequent deep learning model training and prediction, making the entire prediction process more accurate and reliable. This not only helps to promptly identify potential problems with key components, but also provides strong support for engine design optimization, fault diagnosis and maintenance decisions, saving time and cost in many aspects and providing a practical solution for the efficient operation and maintenance of turbojet engines.

[0060] The present invention also proposes the following embodiment for S3-S6:

[0061] S3. For r working conditions, the original flow field data is an n-dimensional vector X (i) (i=1,2,…,r), the temperature field data is X T(i) (i=1,2,…,r), stress field data is X f (i) (i=1,2,…,r), the mean values ​​of temperature field and stress field are:

[0062]

[0063] The sample data of standardized temperature field and stress field are:

[0064]

[0065] The covariance matrix C (n-order square matrix) of the standardized temperature field and stress field data is calculated using the following formula:

[0066]

[0067] Calculate the n eigenvalues ​​of the covariance matrix of the temperature field and the stress field, arrange them in descending order, and take the eigenvector q corresponding to the first m eigenvalues p (1) ,q p (2) ,…,q p (m) , forming an m×n order mapping matrix U p , each eigenvector represents a basic mode in the flow field, and thus the reduced-order flow field data m×r-order matrix Y can be obtained T The fundamental modal coefficient

[0068]

[0069] S4, using the working condition data in the training set and the corresponding temperature field and stress field reduced-order data, the deep learning model is trained to obtain the mapping relationship between the working condition and the reduced-order data;

[0070] Specifically, in this embodiment, the deep learning model used to train the mapping relationship between the operating condition data and the reduced-order data adopts the mamba model, the mean square error is selected as the loss function, and the ADM optimizer is selected as the optimization algorithm. At the same time, it should be noted that two models should be established to train the temperature field model and the stress field model respectively.

[0071] S5, in one embodiment of the present invention, the operating condition data of the temperature field and stress field distribution to be calculated are input into the trained deep learning model to obtain reduced-order temperature field and stress field data;

[0072] Specifically, the operating condition data is input into the temperature field model to predict the fundamental modal coefficients of the temperature field under the corresponding operating condition, and the operating condition data is input into the stress field model to predict the fundamental modal coefficients of the stress field under the corresponding operating condition.

[0073] S6, reduce the order of temperature field and stress field data, and use the transformation relationship between the original flow field data and the reduced order data to recalculate the temperature field and stress field distribution under specific working conditions

[0074] Specifically, in one embodiment of the present invention, the fundamental modal coefficients predicted by the deep learning model are used, and the mapping matrix and the mean value of the flow field are combined to reconstruct the temperature field distribution and the stress field distribution, respectively. The formulas are as follows:

[0075]

[0076] In the formula, U' T with U' f are respectively the pseudo inverse matrices of the mapping matrix obtained in step S3.

[0077] It can be seen that the method for quickly predicting the temperature field and stress field of dangerous parts of key components of a small turbojet engine provided by an embodiment of the present invention adopts a deep learning model and a flow field reduction model, and can realize timely prediction of the temperature field and stress field of local dangerous parts of a small turbojet engine under given working conditions, avoiding the problems of high cost and low efficiency caused by the conventional numerical simulation method in the process of simulating the temperature field and stress field, effectively saving computing cost and time cost, and being able to provide timely temperature field and stress field data feedback to ensure that the engine life is extended and good working performance is maintained.

[0078] A storage medium stores a computer program, which, when executed by a processor, implements a method for quickly predicting the temperature field and stress field of a dangerous part of a key component of a small turbojet engine.

[0079] Specifically, when the computer program stored on the storage medium is executed by the processor, a method for quickly predicting the temperature field and stress field of the dangerous parts of the key components of a small turbojet engine has many beneficial effects. First, the execution of the method is realized through the storage medium, so that the prediction method can be easily deployed and run on different computing devices, which improves the versatility and scalability of the method. The storage medium can store the program for a long time, ensure the stability and sustainability of the method, and provide services for the state monitoring and analysis of the key components of the turbojet engine at any time. Secondly, the method itself has the characteristics of rapid prediction, and can accurately predict the temperature field and stress field in a short time, providing timely decision-making basis for the operation and maintenance of the engine. Through the accurate prediction of the dangerous parts of the key components, potential failure risks can be discovered in advance, reasonable maintenance plans can be formulated, the maintenance cost and downtime of the engine can be reduced, and the reliability and safety of the engine can be improved. In addition, the method covers the whole process from establishing a geometric model to generating a data set, performing data reduction, training a deep learning model, and finally predicting the distribution of the temperature field and stress field, providing engineers with a comprehensive analysis tool, which helps to deeply understand the performance and behavior of the key components of the engine, provide strong support for the design optimization of the engine, and further improve the overall performance and economy of the engine.

[0080] A computer device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for rapidly predicting the temperature field and stress field of dangerous parts of key components of a small turbojet engine as claimed in the claim.

[0081] Specifically, this computer device has significant beneficial effects in realizing a rapid prediction method for the temperature field and stress field of dangerous parts of key components of a small turbojet engine. First, by storing the computer program in the memory, the preservation and repeatability of the prediction method are ensured, so that the method can be conveniently called for prediction analysis at different times and occasions. The processor efficiently executes the program, gives full play to the computing power, and quickly realizes the prediction of the temperature field and stress field of key components of the turbojet engine. This rapid prediction capability provides timely information support for the real-time monitoring and maintenance of the engine, helps to discover potential dangerous parts and failure risks in advance, so as to take corresponding preventive measures and reduce the probability of damage and maintenance costs of the engine. At the same time, the computer device can accurately execute complex prediction methods, covering multiple links from establishing geometric models to data reduction, deep learning model training, etc., providing engineers with accurate analysis tools, helping to deeply understand the performance status of key components of the engine, providing a scientific basis for the design optimization of the engine, and further improving the reliability, safety and economy of the engine. In addition, the flexibility and scalability of the computer device enable it to continuously update and improve the prediction method with the development of technology, adapt to small turbojet engines of different models and specifications, and provide strong technical support for the development of aerospace and other fields.

[0082] The method for rapid prediction of temperature and stress fields of key components of a turbojet engine of the present invention adopts a deep learning model and a flow field reduction model, and can timely predict the temperature and stress fields of local dangerous parts of a turbojet engine under given working conditions, thereby avoiding the problems of high cost and low efficiency caused by the conventional numerical simulation method in the process of simulating the temperature and stress fields, effectively saving computing costs and time costs, and providing timely data feedback to extend the life of the engine and maintain good working performance.

[0083] It should be noted that, in this article, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In addition, "front", "back", "left", "right", "upper" and "lower" in this article are all referenced to the placement state shown in the accompanying drawings.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for rapidly predicting temperature and stress fields at dangerous parts of key components of a small turbojet engine, characterized in that: The method for quickly predicting the temperature field and stress field of the dangerous parts of the key components of a small turbojet engine comprises the following steps: S1. Establish a geometric model for the dangerous parts of a given turbojet engine key component; S2, generating different working condition data and corresponding training data sets and test data sets of temperature fields and stress fields, wherein the generated data sets include temperature field data and stress field data; S3, using the POD flow field reduction model to reduce the temperature field data and the stress field data, and obtain the transformation relationship between the original flow field data and the reduced-order data; S4. Using the working condition data in the training data set and its corresponding temperature field and stress field reduced-order data, the deep learning model is trained to obtain the mapping relationship between the working condition and the temperature field reduced-order data and the stress field reduced-order data respectively; S5. Inputting the working condition data of the temperature field and stress field distribution to be calculated into the trained deep learning model to obtain the reduced-order data of the temperature field and stress field; S6. Recalculate the temperature field and stress field reduced-order data using the transformation relationship between the original flow field data and the reduced-order data to obtain the temperature field and stress field distribution under specific working conditions.

2. The method for rapidly predicting temperature field and stress field of dangerous parts of key components of a small turbojet engine according to claim 1 is characterized in that: In S1, the key components are the parts that are subjected to high temperature, high pressure and complex stress during the operation of the turbojet engine.

3. A method for rapidly predicting temperature and stress fields of dangerous parts of key components of a small turbojet engine according to claim 2, characterized in that: In S2, the operating condition data is selected in a range between the turbojet engine startup state and the extreme operation state.

4. The method for rapidly predicting temperature field and stress field of dangerous parts of key components of a small turbojet engine according to claim 3 is characterized in that: In S2, among the selected operating condition data, the combined operating condition data includes fuel flow, intake pressure and engine speed.

5. The method for rapidly predicting temperature field and stress field of dangerous parts of key components of a small turbojet engine according to claim 4 is characterized in that: In S2, the geometric model of key components of the turbojet engine is also divided using unstructured grids.

6. A method for rapidly predicting temperature and stress fields of dangerous parts of key components of a small turbojet engine according to claim 5, characterized in that: In S2, the temperature field and stress field distribution of key components of the turbojet engine under different working conditions are numerically simulated to obtain temperature field and stress field data of a series of working condition data.

7. A method for rapidly predicting temperature and stress fields of dangerous parts of key components of a small turbojet engine according to claim 6, characterized in that: In S2, the operating condition data and the corresponding temperature field and stress field distribution data are divided into a training data set and a test data set, wherein the training set accounts for 0.

8.

8. The method for rapidly predicting temperature field and stress field of dangerous parts of key components of a small turbojet engine according to claim 7 is characterized in that: In S3, the following steps are also included: S31, respectively obtaining a temperature field average value and a stress field average value for the temperature field data and the stress field data; S32, standardizing the temperature field and stress field data using the obtained temperature field average value and stress field average value; S33, calculating the covariance matrix of the standardized temperature field and stress field data; S34, calculating the eigenvalues ​​of the covariance matrix and sorting them in descending order; S35, taking the eigenvectors corresponding to the top-ranked eigenvalues ​​to form a mapping matrix; S36. By multiplying the temperature field and stress field data matrix on the left by the mapping matrix, low-order temperature field and stress field data, i.e., fundamental modal coefficients of the flow field, are obtained, thereby reducing the order of the temperature field data and the stress field data using the POD flow field reduction model, and obtaining the transformation relationship between the original flow field data and the reduced-order data.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, a method for rapidly predicting temperature fields and stress fields of dangerous parts of key components of a small turbojet engine as described in any one of claims 1 to 8 is implemented.

10. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a method for rapidly predicting temperature fields and stress fields of dangerous parts of key components of a small turbojet engine as described in any one of claims 1 to 8.

Images