A method, device and equipment for predicting heat transfer of engine thermal components

By building a digital twin platform for heat transfer of liquid rocket engines and establishing a heat transfer characteristic model, the problem of the inability to dynamically predict the heat transfer of engine thermal components in the existing technology is solved, and the rapid and accurate prediction of the heat transfer state of the thrust chamber is achieved, and the improvement of the improvement of the heat transfer state of the thrust chamber is achieved. Data analysis efficiency.

CN118467979BActive Publication Date: 2024-11-19XIAN AEROSPACE PROPULSION INST
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410916047.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-09
Publication Date
2024-11-19
Estimated Expiration
2044-07-09

AI Technical Summary

Technical Problem

The prior art cannot conduct real-time dynamic heat transfer prediction of the thermal components of liquid rocket engines, resulting in the inability to accurately predict the heat transfer state of the thrust chamber during test run or flight, affecting the stable operation of the engine.

Method used

By building a digital twin platform for engine heat transfer, multiple sample data are obtained, matrix is ​​established and data preprocessed, network models are trained to establish heat transfer characteristic models, and dynamic heat transfer prediction of engine thermal components is achieved.

Benefits of technology

Real-time dynamic prediction of heat transfer of engine thermal components is realized, and the heat transfer hazard points of the thrust chamber are quickly responded and accurately predicted, which simplifies the analysis process, improves data analysis efficiency, and shortens simulation time.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118467979B_ABST
    Figure CN118467979B_ABST
Patent Text Reader

Abstract

This invention discloses a method, apparatus, and device for predicting heat transfer in engine thermal components, relating to the field of liquid rocket engine technology, to solve the problem that existing technologies cannot perform real-time dynamic prediction of heat transfer in engine thermal components. The method includes: acquiring multiple sample data points of engine thermal components during the heat transfer process from a constructed engine heat transfer digital twin platform, the sample data including at least a first number of gas wall temperature nodes; obtaining a first matrix based on the multiple sample data points and the first number of gas wall temperature nodes; performing data preprocessing on the first matrix to obtain a first orthogonal basis matrix; training a preset network model using the multiple sample data points as input and the first coefficient matrix as output to obtain a heat transfer characteristic model of the engine thermal components; and using this heat transfer characteristic model to perform real-time dynamic prediction of heat transfer in engine thermal components; thus achieving real-time dynamic prediction of heat transfer in engine thermal components.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of liquid rocket engines, and in particular to a method, device and equipment for predicting heat transfer of a thermal component of an engine. Background Art

[0002] At present, the calculation of heat transfer problems of liquid rocket engine thermal components is to analyze the stable working state of the engine. There is no method to predict the variable operating conditions that may be encountered during the engine test or the typical parameter values ​​within the full time range. For example, when calculating the heat transfer problem of the engine thrust chamber, since there are many parameters that need to be changed under stable working conditions, the number of calculations under each condition is large and the time is long. The steady-state analysis method used in the existing technology cannot perform real-time dynamic prediction of the heat transfer of the thrust chamber during actual test or flight.

[0003] Therefore, a more advanced method is needed to predict the heat transfer of engine thermal components in order to solve the problem that the existing technology cannot perform real-time dynamic prediction of the heat transfer of engine thermal components. Summary of the invention

[0004] The object of the present invention is to provide a method, device and equipment for predicting heat transfer of engine thermal components, so as to solve the problem that the prior art cannot perform real-time dynamic prediction of heat transfer of engine thermal components.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] In a first aspect, the present invention provides a method for predicting heat transfer of an engine thermal component, which may include:

[0007] Acquire multiple sample data of the engine thermal component in the heat transfer process from the constructed engine heat transfer digital twin platform, wherein the engine thermal component at least includes the thrust chamber of the engine; the sample data at least includes a first number of air wall temperature nodes;

[0008] Based on the plurality of sample data and the first number of air wall temperature nodes, a first matrix is ​​established;

[0009] Using a preset data processing strategy to perform data preprocessing on the first matrix to obtain a first orthogonal basis matrix;

[0010] Taking the plurality of sample data as input and the first coefficient matrix as output, a preset network model is trained to obtain a heat transfer characteristic model of the engine thermal component; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix;

[0011] Based on the heat transfer characteristic model, dynamic heat transfer prediction is performed on the engine thermal component.

[0012] In a second aspect, the present invention provides a heat transfer prediction device for a thermal component of an engine, which may include:

[0013] An acquisition module, the acquisition module is used to acquire a plurality of sample data of the engine thermal component in the heat transfer process from the constructed engine heat transfer digital twin platform, the engine thermal component at least including the thrust chamber of the engine; the sample data at least includes a first number of air wall temperature nodes;

[0014] An establishing module, the establishing module is used to establish a first matrix based on the plurality of sample data and the first number of air wall temperature nodes;

[0015] An obtaining module, wherein the obtaining module is used to perform data preprocessing on the first matrix using a preset data processing strategy to obtain a first orthogonal basis matrix;

[0016] A training module, the training module is used to take the plurality of sample data as input and the first coefficient matrix as output, to perform model training on a preset network model to obtain a heat transfer characteristic model of an engine thermal component; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix;

[0017] A dynamic prediction module is used to perform dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model.

[0018] In a third aspect, the present invention provides a heat transfer prediction device for an engine thermal component, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the method described in the first aspect is implemented when the processor executes the computer program.

[0019] Compared with the prior art, the present invention provides a method for predicting heat transfer of an engine thermal component, which obtains multiple sample data of the engine thermal component in the heat transfer process from a constructed engine heat transfer digital twin platform; wherein the engine thermal component at least includes the thrust chamber of the engine, and the sample data includes at least a first number of air wall temperature nodes; further based on the multiple sample data and the first number of air wall temperature nodes, a first matrix is ​​established; and a preset data processing strategy is used to perform data preprocessing on the first matrix to obtain a first orthogonal basis matrix; with the multiple sample data as input and the first coefficient matrix as output, a preset network model is trained. A heat transfer characteristic model of the engine thermal component is obtained; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix; finally, dynamic heat transfer prediction is performed on the engine thermal component based on the heat transfer characteristic model; since a corresponding relationship between sample data and the first coefficient matrix is ​​established in the heat transfer characteristic model of the engine thermal component, it can also be understood that a corresponding relationship between each variable in the sample data and the first coefficient matrix is ​​established. Therefore, based on the heat transfer characteristic model of the engine thermal component, dynamic heat transfer prediction is performed on any input variable generated in the heat transfer process of the engine thermal component, thereby realizing real-time dynamic prediction of heat transfer of the engine thermal component. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0021] Figure 1 A first flow chart of a method for predicting heat transfer of an engine thermal component provided by the present invention;

[0022] Figure 2 A schematic diagram of the main architecture of an engine heat transfer digital twin platform provided by the present invention;

[0023] Figure 3 A schematic diagram of the structure of a heat transfer prediction device for a thermal component of an engine provided by the present invention;

[0024] Figure 4 A schematic diagram of the hardware structure of a heat transfer prediction device for an engine thermal component provided by the present invention.

[0025] Figure numerals: 200 - engine heat transfer digital twin platform, 210 - underlying architecture, 220 - functional architecture, 211 - system framework module, 212 - twin display module, 213 - data interaction module, 221 - engine thermal component module, 222 - simulation data management module, 223 - test data management module, 224 - twin prediction module. DETAILED DESCRIPTION

[0026] In order to clearly describe the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, words such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and effects. For example, the first threshold and the second threshold are only used to distinguish different thresholds, and their order is not limited. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not necessarily limit them to be different.

[0027] It should be noted that, in the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0028] In the present invention, "at least one" means one or more, and "plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, the combination of a and b, the combination of a and c, the combination of b and c, or the combination of a, b and c, where a, b, c can be single or multiple.

[0029] With the development of technology, design and manufacturing are constantly moving towards digitalization and intelligence. In the field of liquid rocket engines, the digitization of an engine throughout its entire life cycle, including production, design, manufacturing, testing, flight, and recovery, can preserve its original information to the greatest extent, and can more effectively trace the source of design problems, operating problems, quality problems, etc. that arise during the entire life cycle of the engine.

[0030] At present, the calculation of heat transfer problems of liquid rocket engine thermal components is to analyze the stable working state of the engine. There is no method to predict the variable working conditions that may be encountered during the engine test or the typical parameter values ​​within the full time range. For example: the thrust chamber or generator, an important thermal component in a liquid rocket engine, is mainly used to organize the propellant combustion process and generate high-temperature combustion gas for the normal operation of the engine; because the temperature of the high-temperature combustion gas can be as high as 3000K or more, the thermal environment in the thrust chamber and its heat transfer process are the focus of attention during engine design; and because a single stable working condition requires many parameters to be changed, the calculation will be cumbersome and prone to errors. This requires seeking a faster and simpler method to provide dynamic heat transfer prediction results that change over time.

[0031] Therefore, a more advanced method for dynamically predicting the heat transfer of an engine thermal component is needed to solve the problem in the prior art that the heat transfer of an engine thermal component cannot be dynamically predicted in real time.

[0032] Based on this, the present invention provides a method, device and equipment for predicting heat transfer of engine thermal components, which can dynamically monitor the heat transfer state of heat transfer components during flight or test run, and quickly predict heat transfer danger points of thrust chambers, such as heat transfer components such as thrust chambers; it solves the problem that the existing technology cannot predict the heat transfer of engine thermal components in real time. Next, the technical solution of the present invention is described in detail with reference to the accompanying drawings:

[0033] See also Figure 1 , Figure 1 The first flow chart of a method for predicting heat transfer of an engine thermal component provided by the present invention is shown in FIG. The execution subject is a digital twin platform, a cloud service platform, and a terminal device with simulation capabilities, etc. equipped with the technical solution disclosed in the embodiment of the present invention.

[0034] exist Figure 1 In, the method may include:

[0035] Step 110: Acquire multiple sample data of the engine thermal components during the heat transfer process from the constructed engine heat transfer digital twin platform, wherein the engine thermal components include at least the thrust chamber of the engine; and the sample data include at least a first number of air wall temperature nodes.

[0036] In step 110, firstly, a plurality of sample data of the engine thermal components during the heat transfer process are obtained from the constructed engine heat transfer digital twin platform.

[0037] The criteria for selecting sample points are: the main structural parameters used in the sample point calculation are completely consistent with the actual flight or test structure, and the deviation of the secondary structural parameters from the parameter values ​​in the actual test structure does not exceed 5%. The main structural parameters here are: cylindrical section diameter, throat diameter, outlet diameter, distance between nozzle outlet and injector surface, liquid film starting position, etc.; secondary structural parameters are: number of cooling channels, cooling channel slot width, rib width, slot height, helix angle, etc.

[0038] Specifically, a sample of the heat transfer process of the engine thrust chamber can be selected as an example to illustrate, the samples of the heat transfer process of the thrust chamber can be sorted out, and input conditions and output results can be selected; the main input conditions include but are not limited to chamber pressure P, oxidant flow rate, fuel flow rate, coolant flow rate, liquid film coolant flow rate, etc., and the main output results include gas wall temperature, maximum heat flux density, liquid wall temperature, coolant temperature, etc., carry out sample point calculation, select a certain number of sample points as a training set, and then set aside a small number of sample points as a test set.

[0039] Furthermore, a first number of air wall temperature nodes in each sample data is determined, so that a first matrix can be established based on the plurality of sample data and the first number of air wall temperature nodes.

[0040] As an example, the gas wall temperature Twg is taken as an example to illustrate the implementation process of digital twins.

[0041] According to the sample selection criteria, 100 samples were selected. The main input conditions are chamber pressure P, oxidant flow rate, fuel flow rate, and liquid film coolant flow rate. The output results include 10-20 parameters such as gas wall temperature, maximum heat flux density, liquid wall temperature, coolant temperature, etc. Each output result is discrete on 1000 nodes.

[0042] Step 120: Establishing a first matrix based on the plurality of sample data and the first number of air wall temperature nodes.

[0043] In step 120, according to the example described in step 110, if the plurality of sample data is 100 samples, each operating condition has 4 input variables, and X represents the input matrix, then X is a 4×100 matrix; Y represents the gas wall temperature Twg data (a set of data) obtained by the method of establishing the first matrix, and the gas wall temperature in each sample has 1000 nodes, then Y is a 1000×100 matrix, which can be expressed as Y=(Twg1, Twg2,…, Twg100). The method of establishing the first matrix can be a matrix simulation analysis method in the prior art, which is not specifically limited in the present invention.

[0044] It should be noted that the X matrix in the embodiment of the present invention represents the second matrix, which is the matrix corresponding to the input variables; the Y matrix represents the first matrix, which is the matrix corresponding to the output variables; the X matrix and the Y matrix are both data matrices in the sample data.

[0045] Step 130: Use a preset data processing strategy to perform data preprocessing on the first matrix to obtain a first orthogonal basis matrix.

[0046] Step 140: Using the plurality of sample data as input and the first coefficient matrix as output, a preset network model is trained to obtain a heat transfer characteristic model of the engine thermal component; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix.

[0047] Step 150: Based on the heat transfer characteristic model, dynamic heat transfer prediction is performed on the engine thermal component.

[0048] In step 130 to step 150, the first matrix Y is firstly preprocessed using a preset data processing strategy, and the data preprocessing may include firstly decomposing the first matrix Y to obtain a first orthogonal matrix and a diagonal matrix, and then reducing the order of the first orthogonal matrix to obtain a first orthogonal basis matrix. For example, the first matrix Y may be processed using any one of the singular value decomposition method, the Krylov subspace method or the balanced truncation method to calculate the key modes in the matrix, thereby obtaining a first orthogonal basis matrix; thus, a corresponding relationship between the first matrix and the first orthogonal basis matrix may be established. Because the first matrix includes a corresponding relationship between sample data and a first number of air wall temperature nodes, a corresponding relationship between the sample data and the first matrix is ​​also obtained; further, the inverse moment of the first orthogonal basis matrix may be obtained, and the first coefficient matrix may be obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix. Therefore, the plurality of sample data can be used as input and the first coefficient matrix as output to train the preset network model and obtain the heat transfer characteristic model of the engine thermal component; because the first coefficient matrix, the first orthogonal basis matrix and the first matrix are in a multiplication and division relationship, there is a unique corresponding relationship between the first coefficient matrix and the sample data. There are various input conditions in the sample data, and the corresponding heat transfer prediction data can be obtained by inputting each input condition into the heat transfer characteristic model of the engine thermal component; therefore, based on the heat transfer characteristic model of the engine thermal component obtained after training, it is possible to realize real-time dynamic prediction of heat transfer of different parameter variables under various states of the engine thermal component.

[0049] Based on this, the method disclosed in this specification is used to process the data of the first matrix, extract the data features of the sample library, train the mapping relationship between the input design variables and the characteristic coefficients, realize the prediction of the characteristic coefficients under the variable input conditions, and obtain the remaining parameter information through data reconstruction. After the characteristic model is trained, the output results under the corresponding conditions can be directly output by the characteristic model without the simulation program calculation. Due to the change of this prediction method, the prediction of the heat transfer process of the thrust chamber does not need to go through the heat transfer program calculation and analysis process. The method of the present invention simplifies the method in the prior art into the interpolation of the simulation data results, greatly simplifying the analysis process, so that the dynamic input conditions of the heat transfer process of the heat transfer component can be quickly responded to, and the real-time dynamic prediction of the heat transfer of the engine thermal component can be realized.

[0050] Further, in step 110, the sample data also includes at least a second number of input variables; before establishing a first orthogonal matrix based on multiple sample data and the first number of air wall temperature nodes, it may also include: establishing a second matrix based on multiple sample data and the second number of input variables.

[0051] As an example, taking the thrust chamber as the thermal component, if 100 samples of multiple sample data of the engine thermal component in the heat transfer process are obtained from the constructed engine heat transfer digital twin platform, each working condition has 4 input variables, and X is used to represent the input matrix, then X is a 4×100 matrix; that is, the second matrix X is a 4×100 matrix.

[0052] Further, in step 120, a first matrix is ​​established based on the plurality of sample data and the first number of air wall temperature nodes; specifically:

[0053] Based on the plurality of sample data and the first number of air wall temperature nodes, a first matrix corresponding to the first number of air wall temperature nodes is established.

[0054] As an example, the example in step 110 is used; Y represents the air wall temperature Twg data (a set of data) in the output variable obtained by the traditional simulation analysis method, and each sample has 1000 nodes of the air wall temperature, then Y is a 1000×100 matrix, which can be expressed as Y= (Twg1, Twg2,…, Twg100); that is, the first step of establishing the first matrix Y corresponding to the first number of air wall temperature nodes based on the multiple sample data and the first number of air wall temperature nodes is completed.

[0055] In step 130, the method of using a preset data processing strategy to perform data preprocessing on the first matrix to obtain a first orthogonal basis matrix may include the following two steps:

[0056] Step 1: Use the POD method to extract the data features of the sample library, decompose the first matrix Y, and use the singular value decomposition method to obtain the first orthogonal matrix U1.

[0057] Specifically, the first matrix Y is decomposed by using a singular value decomposition method, and the formula may be used:

[0058] (1)

[0059] Decompose the first matrix Y; U1 represents a 1000×100 first orthogonal matrix, S1 represents a 1000×100 diagonal matrix, and V1 represents a 100×100 second orthogonal matrix. The elements on the diagonal line of the diagonal matrix are called singular values.

[0060] The second step is to calculate the first orthogonal basis matrix based on the diagonal matrix and the first orthogonal matrix using a preset function calculation reduction strategy.

[0061] Specifically, based on the diagonal matrix and the first orthogonal matrix, a preset function calculation reduction strategy can be used to calculate the first orthogonal basis matrix, which can include: based on the diagonal matrix, determining each singular value in the diagonal matrix; the singular value represents the element on the diagonal line of the diagonal matrix; according to each singular value in the diagonal matrix, the formula can be used:

[0062] (2)

[0063] The energy corresponding to each basis vector in the first orthogonal basis matrix U1 is calculated; where: i represents the energy corresponding to each orthogonal basis vector in the first orthogonal matrix, λ i represents the i-th singular value, N represents the first preset limit value, and j represents a value from 1 to N.

[0064] When sum(w 1:r )<N, the matrix corresponding to columns 1 to r in the first orthogonal matrix U1 is determined as the first orthogonal basis matrix U r .

[0065] The inverse matrix of the first orthogonal basis matrix can be further calculated , multiply the inverse matrix of the first orthogonal basis matrix by the first matrix Y to obtain the first coefficient matrix A; thus, the preset network model can be trained with multiple sample data as input and the first coefficient matrix as output, thereby obtaining the heat transfer characteristic model of the engine thermal component.

[0066] It should be noted that this specification does not limit the number of samples, which can be 100 sample data, or 200 or other numbers of sample data; nor does it limit the method of model training, which can be self-learning training or semi-supervised learning training.

[0067] As an example, the example in step 120 is used to explain in detail by taking the engine thrust chamber as an example of the heat transfer component.

[0068] First, a preset data processing strategy is used to decompose the first matrix to obtain a first orthogonal matrix; then the first orthogonal matrix is ​​reduced in order, and the key modes in the first orthogonal matrix are selected to obtain a first orthogonal basis matrix U r , and multiply the inverse matrix of the first orthogonal basis matrix by the first matrix to obtain a first coefficient matrix; wherein r represents the number of orthogonal basis vector columns, and A is the first coefficient matrix of orthogonal basis vectors.

[0069] The method of taking r is: describe each singular value in the S1 matrix as λ i , using formula (2) to calculate the first orthogonal matrix U r The energy corresponding to each basis vector in .

[0070] Further, according to the preset limit adjustment, when the energy corresponding to the sum of the current r columns of orthogonal basis vectors reaches the first limit N, it is considered that these r columns can be used to reflect the energy value of the original matrix, and the data of these r columns are taken as the effective data. That is, when sum(w 1:r )<N, the matrix corresponding to columns 1 to r in the first orthogonal matrix is ​​determined as the first orthogonal base matrix.

[0071] Each column coefficient can be considered to be generated by the change of the sample point. Assuming that in the data characteristics of the air wall temperature, r=8 is obtained, the first orthogonal basis matrix U can be selected from the first orthogonal matrix U1 r , the matrix is ​​a 1000×8 matrix, so the original 1000-dimensional data matrix corresponding to each sample point is compressed into an 8-dimensional data matrix.

[0072] Based on the above method, the data preprocessing of the first matrix is ​​completed, and the first orthogonal basis matrix is ​​obtained; it should be noted that the above method is illustrated by taking the thrust chamber as an example of the thermal component of the engine, and the method described in this specification is also applicable to other thermal components of the engine. The present invention takes the thrust chamber as an example, and mainly uses the verified thrust chamber simulation database as the basic data source. By decomposing and reducing the order of the database, such as using the POD method in the orthogonal decomposition method for order reduction, the data features of the sample library are extracted; the data reduction method can also be the Krylov subspace method or the balanced truncation method. Further train the mapping relationship between the input design variables and the characteristic coefficients, and finally realize the prediction of the characteristic coefficients under variable input conditions, and obtain the remaining parameter information through data reconstruction; compared with the methods in the prior art, the present invention also reduces the amount of data analysis and greatly improves the efficiency of data analysis.

[0073] It should be noted that by using the method described in the present invention to decompose and reduce the order of data, the simulation time of tens of minutes can be shortened to within one minute, thereby quickly improving the work efficiency of data analysis.

[0074] Further, the formula can be used:

[0075] (3)

[0076] represents the correspondence between the first matrix, the first orthogonal basis matrix and the first coefficient matrix. r represents the first orthogonal basis matrix, and A represents the first coefficient matrix.

[0077] Similarly, formula (3) can also be expressed as:

[0078] (4)

[0079] For the first matrix Y and the first orthogonal basis matrix U r The first coefficient matrix A can be obtained by converting the corresponding relationship between the matrix A (8×100 matrix) and the Y matrix. Therefore, the unique corresponding relationship between the matrix A (8×100 matrix) and the input X (4×100 matrix) is established. Since the Y matrix has a corresponding relationship with the X matrix, the unique corresponding relationship between the matrix A (8×100 matrix) and the input X (4×100 matrix) is also established. represents the inverse matrix of the first orthogonal basis matrix.

[0080] Therefore, the following formula can be used

[0081] Y new = U r ×A new (5)

[0082] Get the new physical field; where Ynew Represents the new physical field (Y new For the input new sample point X new There is a unique orthogonal matrix), U r represents the first orthogonal basis matrix, A new Represents the new coefficient matrix.

[0083] Furthermore, new prediction results can be obtained based on the new physical field.

[0084] Furthermore, before step 140, it may also include: using the test set data to judge the prediction accuracy of the heat transfer characteristic model of the engine thermal component, to judge whether the prediction accuracy meets the preset requirements, if so, it can be used, if not, it needs to further optimize the preset restrictions.

[0085] Specifically, before performing dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model, the method may further include:

[0086] Acquire multiple test set data; based on the multiple test set data, determine whether the heat transfer prediction error of the heat transfer characteristic model is greater than a first error threshold; if the heat transfer prediction error of the heat transfer characteristic model is less than or equal to the first error threshold, perform dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model; wherein the first error threshold can be calculated based on historical data to obtain an error threshold that meets the design requirements, or a prediction threshold can be set based on the latest test requirements; in the present invention, the first error threshold can be set to 5%.

[0087] Furthermore, if the heat transfer prediction error of the heat transfer characteristic model is greater than the first error threshold, a second preset limit is set; the method of step 130 to step 140 can be used to first pre-process the first matrix based on the second preset limit and the first matrix using a preset data processing strategy to obtain a second orthogonal basis matrix; the inverse matrix of the second orthogonal basis matrix is ​​multiplied by the first matrix to obtain a second coefficient matrix. Then, with multiple sample data as input and the second coefficient matrix as output, the preset network model is trained until the heat transfer prediction error of the heat transfer characteristic model is less than or equal to the first error threshold, and the heat transfer characteristic model of the engine thermal component is obtained.

[0088] After confirming that the prediction error of the heat transfer characteristic model of the engine thermal component is less than or equal to the first error threshold, a dynamic heat transfer prediction is performed on the engine thermal component based on the heat transfer characteristic model.

[0089] Further, in step 150, dynamic heat transfer prediction of the engine thermal component is performed based on the heat transfer characteristic model, which may include: real-time prediction. After obtaining the accurate heat transfer characteristic model, a time dimension t is added to the calculation process, and the main input conditions are plotted as a curve changing with time, and discretized according to time. Every moment can be predicted in real time according to the method disclosed in the present invention. Every moment can generate a data list identical to the sample point, and output important data values ​​in the predicted data, such as the maximum value of the gas wall temperature max (Twg) at a certain moment and the node position i, the maximum value of the heat flux density and the position of occurrence, etc., and this value is plotted as a curve changing with time and displayed in the display window, that is, the real-time prediction of the main heat transfer data of the thrust chamber is completed. This real-time prediction can reduce the time for working condition prediction from hours to seconds.

[0090] Further, taking the thrust chamber as an example, at the beginning of the design of the liquid rocket engine, it is necessary to design and preliminarily calculate the internal surface of the thrust chamber, the parameters of the cooling groove, the combustion pressure, and the mixing ratio. Then, according to the location of the heat transfer danger point and the temperature exceeding the material limit, the design parameters of the thrust chamber are modified and improved, and even the structure or design parameters are changed. At this time, due to the changes in the structure and design parameters of the thrust chamber, the heat transfer calculation program used for analysis will generate multiple versions; in the subsequent engine test and operation process, the structural parameters are determined, and the test and operation parameters will change within a certain range. The simulation of the test process will generate multiple versions of the heat transfer program, and managing the changes and change states of these programs brings great difficulties to the designers. In addition, although the data that can be measured during the test is limited, the data obtained from the test is very valuable. These data are very meaningful for the correction and improvement of the heat transfer model, so it is also necessary to centrally manage them. However, there is no complete data management method in the prior art to effectively manage the programs and test data during the design, test, and operation process. Therefore, a more advanced data management method is needed to effectively manage the design parameters, simulation programs, data and test data generated during design and testing.

[0091] Based on this, the present invention also establishes an engine heat transfer digital twin platform including rocket engine thermal components, that is, before step 110, it can also include building an engine heat transfer digital twin platform, using the engine heat transfer digital twin platform to effectively manage the design parameters, simulation programs, data and test data generated in the design and test; so that through Figure 1 The prediction method shown can establish a heat transfer characteristic model of the engine thermal components based on the digital twin platform and perform real-time dynamic heat transfer prediction of the engine thermal components; its prediction accuracy is higher and the prediction efficiency will also be higher.

[0092] For example: During the engine test or flight, although there are preset engine input conditions and operating conditions, in the actual test work, various deviations will always occur, making the input conditions slightly different from the conditions expected to be used in the original test, thus causing engine damage caused by heat transfer hazards during the test. The digital twin of the heat transfer process is to connect the digital twin platform with the test system or flight telemetry system in real time, obtain the main conditions of the engine thrust chamber in real time, and predict the transient process in real time. The prediction results can be expressed in real time in the form of curves on the digital twin platform, and the working state of the thrust chamber can be obtained more intuitively. When the temperature or heat flux density of the thrust chamber abnormally rises to a certain threshold, the input conditions (pressure, flow and other parameters) are adjusted in time to ensure that the temperature or heat flux density of the thrust chamber is within the normal range. This adjustment method can be instantly fed back to the engine being tested or flying for correcting the working state. In addition, the digital twin process can also be used for the health management of thrust chamber heat transfer.

[0093] Preferably, see Figure 2 , Figure 2 A schematic diagram of the main architecture of an engine heat transfer digital twin platform provided by the present invention.

[0094] exist Figure 2 In the embodiment, the engine heat transfer digital twin platform 200 includes at least an underlying architecture 210 and a functional architecture 220. The engine heat transfer digital twin platform 200 can effectively manage the real model, simulation program, simulation results, and test data in a unified manner, and can effectively consolidate and integrate existing computing experience, so that the engine thrust chamber heat transfer design and management work has inheritance and extensibility.

[0095] Specifically, the underlying architecture 210 includes at least a system framework module 211, a twin display module 212, and a data interaction module 213; the functions of each module in the underlying architecture 210 are as follows:

[0096] The system framework module 211 is mainly used to provide underlying support for the development of other business functions, which is mainly reflected in the startup of the engine heat transfer platform and system, the rapid loading of data of each functional module and the process operation.

[0097] The twin display module 212 is mainly used to provide an easy-to-operate and convenient user interaction interface, allowing users to intuitively operate and run the heat transfer twin platform system. At the same time, it integrates and processes the important data generated, and provides users with a vivid and technological display effect, allowing users to more easily see the engine thrust chamber structure, important parameters, input conditions, and key temperatures and temperature distribution in the heat transfer process, as well as their changes over time.

[0098] The data interaction module 213 is mainly used in the background to provide big data management and interaction, using an open source relational database, integrated with the system through the JDBC / ODBC standard interface, and providing data storage, management, and query services for the system.

[0099] The functional architecture 220 at least includes an engine thermal component module 221 (real model module), a simulation data management module 222, a test data management module 223, and a twin prediction module 224; the functions of each module in the functional architecture 220 are as follows:

[0100] The engine thermal component module 221 is mainly used to store the structural parameters, working parameters and physical parameters of the engine thermal component; taking the thrust chamber as an example, it mainly stores the thrust chamber structural parameters, such as: thrust chamber inner surface, regenerative cooling groove width, rib width, rib height, number of grooves, whether spiral grooves, downstream or reverse flow, cooling flow path and other conditions required for regenerative cooling, structural outer surface required for radiation heat exchange, etc.; working parameters, such as: chamber pressure, oxidant flow rate, fuel flow rate, liquid film flow rate, liquid film position, etc.; physical parameters, such as: oxidant physical properties, fuel physical properties, combustion product physical properties, liquid film physical properties, structural wall physical properties, regenerative coolant parameters required for regenerative cooling, and nozzle outer surface emissivity required for radiation heat exchange. The physical parameters required for heat transfer are generally: pressure, temperature, density, specific heat capacity, thermal conductivity, viscosity coefficient, vaporization latent heat, etc. Furthermore, parameter management and external model calling are adopted. For important parameters, parameters of different versions of models are recorded in the form of data charts, and models with large changes are marked and annotated.

[0101] It should be noted that for structures that are difficult to describe, such as the thrust chamber profile and cooling flow path, they can be recorded using three-dimensional modeling software (such as NX, PROE, etc.) and called and viewed using a data interface.

[0102] The simulation data management module 222 is mainly used to store the simulation model, code and simulation calculation results of the heat transfer process of the engine thermal component; and to control the simulation model of the heat transfer process of the engine thermal component to perform steady-state prediction of the heat transfer state of the engine thermal component, and to solidify the program and method of heat transfer analysis. For example, the simulation model, code and simulation calculation results of the heat transfer process of the thrust chamber are stored and controlled.

[0103] The simulation model is mainly used for design and prediction. In the early stage of engine design, when there is no physical model, the steady-state prediction of the heat transfer state of the thrust chamber can be carried out through the parameterized simulation model. It can be further improved later. The integration into the simulation model can solidify the procedures and methods of heat transfer analysis. The purpose of summarizing the simulation results is mainly to form a result database, which is convenient for training feature models when the heat transfer data is twinned. Main implementation method: Since the format of the parameters and their values ​​output by the simulation analysis is relatively fixed, it is relatively convenient to implement management, mainly to ensure the one-to-one correspondence and perfection of the input parameters and output parameters.

[0104] The test data management module 223 is mainly used to store the heat transfer data during the engine test run. In the actual engine development process, the temperature of the engine thrust chamber will be measured, and the radiation cooling thrust chamber can generally obtain the actual temperature value of the outer wall of the thrust chamber. However, for the regenerative cooling engine, due to its complex structure, it is impossible to directly measure the inner wall temperature and heat flux density, and it will be selected to reflect the heat flux level of the engine thrust chamber by measuring the temperature rise of the regenerative coolant. At the same time, as an important component of the rocket engine, its actual test run sample data is relatively small, and it must be effectively utilized and managed. At this time, by collecting the test run data to form a database, on the one hand, the thrust chamber heat transfer simulation model can be verified, and on the other hand, the test run data can be uniformly managed, and this heat transfer data system can retain an interface with the engine test run data system for data intercommunication and update. Main implementation method: At present, the thrust chamber heat transfer data that can be obtained by test analysis is limited. Generally, the effective data is the temperature rise of the regenerative coolant in the thrust chamber. This temperature rise data can be used as effective data for simulation model verification. During each test, the engine must test different working conditions and the working conditions are constantly changing. Therefore, the test data can be decomposed and managed for more effective use.

[0105] The twin prediction module 224 is mainly used to make real-time predictions of the heat transfer state of the engine. The twin prediction module includes a characteristic model training unit and a heat transfer real-time prediction unit. The characteristic model training unit is used to perform characteristic model training to obtain the heat transfer characteristic model of the engine thermal component; the heat transfer real-time prediction unit performs dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model, and makes real-time predictions of heat transfer danger points; the twin prediction module 224 can shorten the simulation time of tens of minutes to within one minute, quickly improving work efficiency.

[0106] It should be noted that the training method and prediction method of the twin prediction module 224 are consistent with the methods in steps 110 to 150, and will not be repeated here. Of course, the twin prediction module 224 can also continuously supplement and add simulation results or test data to update the feature model, thereby improving the accuracy of the twin data.

[0107] Specifically, the result obtained by heat transfer simulation is a steady-state heat transfer analysis at each time point, while the real-time prediction is a dynamic prediction. After a quick prediction at each moment, typical data can be extracted to form a curve or icon for the input conditions changing over time. This dynamic real-time prediction can predict the iconic heat transfer danger points. The twin prediction module 224 is characterized in that a feature model is trained on the basis of simulation data. In the process of forming the feature model, the data is re-interpolated and called according to different thrust chamber input conditions, and typical heat transfer data is quickly given to determine whether the heat transfer state of the engine thrust chamber under this time sequence is within the safety threshold; the typical heat transfer data can be the maximum air wall temperature, the maximum heat flux density, the regeneration coolant outlet temperature, etc.

[0108] Taking the thrust chamber as an example, its implementation method is: mainly using the verified thrust chamber simulation database as the basic data source, and reducing the order of the database; there are many methods for reducing the order, such as Krylov subspace method, balanced truncation method or orthogonal decomposition method, etc.; the present invention uses the POD method in the orthogonal decomposition method for order reduction, extracts the data features of the sample library, trains the mapping relationship between the input design variables and the characteristic coefficients, and finally realizes the prediction of the characteristic coefficients under variable input conditions, and obtains the remaining parameter information through data reconstruction. After the characteristic model is trained, the results under the corresponding conditions can be directly output by the characteristic model without the calculation of the simulation program. Due to the change of this prediction method, the prediction of the heat transfer process of the thrust chamber eliminates the process of calculation and analysis of the heat transfer program, and is simplified to the interpolation of the simulation data results, which greatly simplifies the analysis process, thereby shortening the prediction speed of a single state, so that the dynamic input conditions of the heat transfer process of the thrust chamber can be quickly responded to and predicted, and the digital twin of the heat transfer process of the thrust chamber of the liquid rocket engine can be realized.

[0109] Based on this, the present invention adopts a heat transfer prediction method for an engine thermal component provided in the first aspect to establish a rocket engine heat transfer digital twin platform, which can complete the management of the real model of the engine thermal component, the heat transfer simulation model and program, the heat transfer analysis results, and the heat transfer data during the test or flight; on the basis of the digital twin platform, the thrust chamber of the thermal component is taken as the object, and an example is given to solve the dynamic real-time monitoring of the heat transfer work of the thrust chamber during the flight or test, and quickly predict the heat transfer danger points of the thrust chamber; therefore, the heat transfer characteristic model of the engine thermal component obtained by the method described in the first aspect of the present invention can perform real-time dynamic heat transfer prediction on any input variable generated in the heat transfer process of the engine thermal component, thereby realizing real-time dynamic prediction of the heat transfer of the engine thermal component; it efficiently solves the problem that the existing technology cannot perform real-time dynamic prediction of the heat transfer of the engine thermal component.

[0110] In the second aspect, based on the same technical concept as the first aspect, the present invention provides a heat transfer prediction device for an engine thermal component. Figure 3 , Figure 3 A schematic structural diagram of a heat transfer prediction device for an engine thermal component provided by the present invention.

[0111] exist Figure 3 In the embodiment, the device may include:

[0112] An acquisition module 310 is used to acquire a plurality of sample data of an engine thermal component in a heat transfer process from the constructed engine heat transfer digital twin platform, wherein the engine thermal component at least includes a thrust chamber of the engine; and the sample data at least includes a first number of air wall temperature nodes;

[0113] An establishing module 320, configured to establish a first matrix based on the plurality of sample data and the first number of air wall temperature nodes;

[0114] An obtaining module 330 is used to perform data preprocessing on the first matrix using a preset data processing strategy to obtain a first orthogonal basis matrix;

[0115] A training module 340 is used to perform model training on a preset network model using the plurality of sample data as input and a first coefficient matrix as output to obtain a heat transfer characteristic model of an engine thermal component; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix;

[0116] The dynamic prediction module 350 is used to perform dynamic heat transfer prediction on the engine thermal components based on the heat transfer characteristic model.

[0117] The present invention proposes a heat transfer prediction device for an engine thermal component, which obtains multiple sample data of the engine thermal component in the heat transfer process from a constructed engine heat transfer digital twin platform through an acquisition module, wherein the engine thermal component at least includes the thrust chamber of the engine, and the sample data includes at least a first number of air wall temperature nodes; further based on the establishment module, based on the multiple sample data and the first number of air wall temperature nodes, a first matrix is ​​established; and based on the acquisition module, a preset data processing strategy is used to perform data preprocessing on the first orthogonal matrix to obtain a first orthogonal base matrix; then based on the training module, with the multiple sample data as input and the first coefficient matrix as output, a preset network model is trained to obtain a heat transfer characteristic model of the engine thermal component; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal base matrix by the first matrix; finally, a dynamic prediction module is used to perform dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model; real-time dynamic prediction of heat transfer of the engine thermal component is realized.

[0118] Third, see Figure 4 , Figure 4 A schematic diagram of the hardware structure of a heat transfer prediction device for an engine thermal component provided by the present invention.

[0119] exist Figure 4 In the embodiment, the device 400 is a heat transfer prediction device for an engine thermal component. The device 400 includes a memory 420, a first processor 410, a second processor 450, and a computer program stored in the memory 420 and executable on the first processor 410 and / or the second processor 450. When the first processor 410 and / or the second processor 450 executes the computer program, the method described in the first aspect, a heat transfer prediction method for an engine thermal component, is implemented.

[0120] Optionally, the device 400 may include a communication interface 430. There may be one or more communication interfaces 430. The communication interface 430 may use any transceiver or other device for communicating with other devices or communication networks.

[0121] Optionally, the device 400 may further include a communication line 440. The communication line 440 may include a path to transmit information between the above components.

[0122] It should be noted that the memory may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may exist independently and be connected to the processor through a communication line. The memory may also be integrated with the processor.

[0123] Optionally, the computer-executable instructions in the embodiment of the present invention may also be referred to as application program codes, which is not specifically limited in the embodiment of the present invention.

[0124] In a specific implementation, as an example, Figure 4 As shown, the first processor 410 and the second processor 450 may include one or more CPUs, such as Figure 4 CPU0 and CPU1 in.

[0125] Although the present invention is described herein in conjunction with various embodiments, in the process of implementing the claimed invention, those skilled in the art may understand and implement other variations of the disclosed embodiments by viewing the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "one" or "an" does not exclude multiple situations. A single processor or other unit may implement several functions listed in a claim. Certain measures are recorded in different dependent claims, but this does not mean that these measures cannot be combined to produce good results.

[0126] Although the present invention has been described in conjunction with specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present invention. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present invention as defined by the appended claims and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present invention. Obviously, those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, the present invention is intended to include such modifications and variations if they fall within the scope of the claims of the present invention and their equivalents.

Claims

1. A method for predicting heat transfer of an engine thermal component, characterized in that: include: Acquire multiple sample data of the engine thermal components during the heat transfer process from the constructed engine heat transfer digital twin platform, wherein the engine thermal components at least include the thrust chamber of the engine; The sample data includes at least a first number of air wall temperature nodes; the engine heat transfer digital twin platform includes at least an underlying architecture and a functional architecture; the underlying architecture includes at least a system framework module, a twin display module and a data interaction module; the functional architecture includes at least an engine thermal component module, a simulation data management module, a test data management module and a twin prediction module; wherein the engine thermal component module is used to store the structural parameters, working parameters and physical parameters of the engine thermal component; the simulation data management module is used to store the simulation model, code and simulation calculation results of the heat transfer process of the engine thermal component; the test data management module is used to store the heat transfer data during the engine test; the twin prediction module is used to make real-time predictions on the heat transfer state of the engine thermal component; Based on the plurality of sample data and the first number of air wall temperature nodes, a first matrix is ​​established; Using a preset data processing strategy to perform data preprocessing on the first matrix to obtain a first orthogonal basis matrix; Taking the plurality of sample data as input and the first coefficient matrix as output, a preset network model is trained to obtain a heat transfer characteristic model of the engine thermal component; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix; Based on the heat transfer characteristic model, dynamic heat transfer prediction is performed on the engine thermal component.

2. The method for predicting heat transfer of an engine thermal component according to claim 1, characterized in that: The step of using a preset data processing strategy to perform data preprocessing on the first matrix to obtain a first orthogonal basis matrix includes: Decomposing the first matrix by singular value decomposition to obtain a diagonal matrix and a first orthogonal matrix; Based on the diagonal matrix and the first orthogonal matrix, a preset function calculation reduction strategy is adopted to obtain the first orthogonal basis matrix.

3. The method for predicting heat transfer of an engine thermal component according to claim 2, characterized in that: The step of obtaining the first orthogonal basis matrix by using a preset function calculation reduction strategy based on the diagonal matrix and the first orthogonal matrix includes: Based on the diagonal matrix, determining each singular value in the diagonal matrix; the singular value represents an element on a diagonal line in the diagonal matrix; According to each singular value in the diagonal matrix, the formula is used: ; The energy corresponding to each basis vector in the first orthogonal matrix is ​​calculated; wherein: w i represents the energy corresponding to each orthogonal basis vector in the first orthogonal matrix, λ i represents the i-th singular value, N represents the first preset limit value, and j represents a value from 1 to N; When sum(w 1:r )<N, the matrix corresponding to columns 1 to r in the first orthogonal matrix is ​​determined as the first orthogonal base matrix.

4. The method for predicting heat transfer of an engine thermal component according to claim 1, characterized in that: The step of performing dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model comprises: Get multiple test set data; Based on the plurality of test set data, determining whether the heat transfer prediction error of the heat transfer characteristic model is greater than a first error threshold; If the heat transfer prediction error of the heat transfer characteristic model is less than or equal to a first error threshold, then performing a dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model; If the heat transfer prediction error of the heat transfer characteristic model is greater than the first error threshold, setting a second preset limit value; Based on the second preset limit value and the first matrix, using a preset data processing strategy to perform data preprocessing on the first matrix to obtain a second orthogonal basis matrix; Taking the plurality of sample data as input and the second coefficient matrix as output, the preset network model is trained until the heat transfer prediction error of the heat transfer characteristic model is less than or equal to the first error threshold, thereby obtaining the heat transfer characteristic model of the engine thermal component; the second coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the second orthogonal basis matrix by the first matrix; Based on the heat transfer characteristic model, dynamic heat transfer prediction is performed on the engine thermal component.

5. The method for predicting heat transfer of an engine thermal component according to claim 1, characterized in that: The sample data also includes at least a second number of input variables; The step of establishing a first matrix based on the plurality of sample data and the first number of air wall temperature nodes includes: establishing a second matrix based on the plurality of sample data and the second number of input variables.

6. The method for predicting heat transfer of an engine thermal component according to claim 1, characterized in that: The simulation data management module is also used to control the simulation model of the heat transfer process of the engine thermal component to perform steady-state prediction of the heat transfer state of the engine thermal component, and to solidify the program and method of heat transfer analysis.

7. The method for predicting heat transfer of an engine thermal component according to claim 6, characterized in that: The twin prediction module at least includes a characteristic model training unit and a heat transfer actual situation prediction unit; The characteristic model training unit is used to perform characteristic model training to obtain a heat transfer characteristic model of the engine thermal component; The heat transfer real-time prediction unit performs dynamic heat transfer prediction on the engine thermal components based on the heat transfer characteristic model, and performs real-time forecasting on heat transfer danger points.

8. A heat transfer prediction device for an engine thermal component, characterized in that: include: An acquisition module, the acquisition module is used to acquire a plurality of sample data of the engine thermal components in the heat transfer process from the constructed engine heat transfer digital twin platform, the engine thermal components at least including the thrust chamber of the engine; The sample data includes at least a first number of air wall temperature nodes; the engine heat transfer digital twin platform includes at least an underlying architecture and a functional architecture; the underlying architecture includes at least a system framework module, a twin display module and a data interaction module; the functional architecture includes at least an engine thermal component module, a simulation data management module, a test data management module and a twin prediction module; wherein the engine thermal component module is used to store the structural parameters, working parameters and physical parameters of the engine thermal component; the simulation data management module is used to store the simulation model, code and simulation calculation results of the heat transfer process of the engine thermal component; the test data management module is used to store the heat transfer data during the engine test; the twin prediction module is used to make real-time predictions on the heat transfer state of the engine thermal component; An establishing module, the establishing module is used to establish a first matrix based on the plurality of sample data and the first number of air wall temperature nodes; An obtaining module, wherein the obtaining module is used to perform data preprocessing on the first matrix using a preset data processing strategy to obtain a first orthogonal basis matrix; A training module, the training module is used to take the plurality of sample data as input and the first coefficient matrix as output, to perform model training on a preset network model to obtain a heat transfer characteristic model of an engine thermal component; the first coefficient matrix is ​​a matrix obtained by multiplying the inverse matrix of the first orthogonal basis matrix by the first matrix; A dynamic prediction module is used to perform dynamic heat transfer prediction on the engine thermal component based on the heat transfer characteristic model.

9. A heat transfer prediction device for an engine thermal component, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.