Aero-engine digital twin modeling optimization method based on transfer learning

Abstract

The invention relates to the technical field of modeling, in particular to an aero-engine digital twin modeling optimization method based on transfer learning. According to the method, the first digital twinborn model is established by combining the environment characteristic data and the control characteristic data through the random forest algorithm, so that the performance and behavior changes of the aero-engine under different working conditions can be captured, and fine tuning and optimization can be performed according to real-time data through transfer learning of the second digital twinborn model; the third digital twin model can adapt to data distribution changes and offsets possibly existing in a target task through transfer learning, transfer learning of source domain data is beneficial to enabling the model to be more stable, the changes can be effectively coped with, and on the basis of transfer learning data of the third digital twin model, the generalization ability and adaptability of the model are improved. The parameters of the second digital twinning model can be finely adjusted and optimized, and the optimization not only can improve the prediction precision of the model, but also is helpful to improve the generalization ability of the model on new data.

Description

Technical Field

[0001] The present invention relates to the field of modeling technology, and in particular to an aircraft engine digital twin modeling optimization method based on transfer learning. Background Art

[0002] Transfer learning is an algorithm that uses the knowledge gained from the source domain in solving the task to improve the target task given the source domain and source task, target domain and target task.

[0003] The aircraft engine is one of the core components of the aircraft, responsible for providing power to propel the aircraft to fly. The digital twin modeling of the aircraft engine is to build a simulation model of the real engine in the physical space in the virtual space. By simulating and analyzing the operating status of the aircraft engine, it can help engineers optimize the design, monitor the operation, and perform maintenance. At present, the digital twin modeling of aircraft engines is modeled through sample data, but there is a problem that the model accuracy is not enough due to the insufficient number of sample data, which makes the output exhaust temperature parameters not accurate enough. Summary of the invention

[0004] The main purpose of the present invention is to provide an aircraft engine digital twin modeling optimization method based on transfer learning, aiming to solve the technical problems in the prior art.

[0005] The present invention proposes an aero-engine digital twin modeling optimization method based on transfer learning, comprising:

[0006] Acquire historical target data and multiple real-time target data of the aircraft engine, wherein the historical target data includes environmental characteristic data and control characteristic data;

[0007] Using a random forest algorithm to construct a first digital twin model according to the environmental characteristic data and the control characteristic data;

[0008] Training the first digital twin model according to the plurality of real-time target data to obtain a second digital twin model;

[0009] Obtain source domain data related to historical target data;

[0010] Selecting a transfer learning model according to the source domain data, and inputting the source domain data into the transfer learning model for training to obtain a third digital twin model;

[0011] Acquire transfer learning data between the third digital twin model and the second digital twin model, and optimize the second digital twin model according to the transfer learning data to obtain a digital twin optimization model.

[0012] Preferably, the step of using a random forest algorithm to construct a first digital twin model according to the environmental feature data and the control feature data comprises:

[0013] Acquire atmospheric density data according to the environmental characteristic data, and acquire fuel flow data according to the control characteristic data;

[0014] Obtaining a cross-sectional area of ​​an air inlet of an aircraft engine, an air temperature of the air inlet, a first airflow total pressure of the air inlet, and a second airflow total pressure of the exhaust port;

[0015] A first digital twin model is constructed according to the cross-sectional area, the intake temperature, the first airflow total pressure, the second airflow total pressure, the fuel flow data, and the atmospheric density data, wherein the first digital twin model is:

[0016]

[0017] Among them, P(W) represents the predicted exhaust temperature, J(W) represents the intake temperature, R(L) represents the fuel flow data, J(M) represents the cross-sectional area, QY(1) represents the total pressure of the first airflow, QY(2) represents the total pressure of the second airflow, k represents the first empirical weight coefficient, n represents the second empirical weight coefficient, and m represents the third empirical weight coefficient.

[0018] Preferably, the step of acquiring the atmospheric density data according to the environmental characteristic data and acquiring the fuel flow rate data according to the control characteristic data comprises:

[0019] Acquire temperature parameters, pressure parameters and altitude parameters according to the environmental characteristic data;

[0020] Obtain the temperature gradient and gas constant of the environment in which the aircraft engine is located;

[0021] The atmospheric density data is calculated according to the temperature parameter, pressure parameter, altitude parameter, temperature gradient and gas constant, wherein the calculation formula is:

[0022]

[0023] Among them, Q(N) represents atmospheric density data, Y(C) represents pressure parameters, W(C) represents temperature parameters, Q(C) represents gas constant, W(T) represents temperature gradient, and H(C) represents altitude parameters;

[0024] Acquire a throttle lever angle parameter, a flight speed parameter, and a throttle opening parameter according to the control characteristic data;

[0025] Obtain basic fuel flow for aircraft engines;

[0026] The fuel flow data is calculated according to the basic fuel flow, the throttle lever angle parameter, the flight speed parameter and the throttle opening parameter, wherein the calculation formula is:

[0027] R(L)=J(R)*[a*Y(J)+b*F(S)+c*Y(K)];

[0028] Wherein, R(L) represents the fuel flow data, J(R) represents the basic fuel flow, a represents the weight of the throttle lever angle parameter, Y(J) represents the throttle lever angle parameter, b represents the weight of the flight speed parameter, F(S) represents the flight speed parameter, c represents the weight of the throttle opening parameter, and Y(K) represents the throttle opening parameter.

[0029] Preferably, the step of training the first digital twin model according to the plurality of real-time target data to obtain the second digital twin model comprises:

[0030] Select any real-time target data as a training set and input it into the first digital twin model for training, and input the remaining real-time target data as a verification set into the first digital twin model for verification, to obtain multiple training results;

[0031] The model parameters of the first digital twin model are adjusted according to the multiple training results to obtain the second digital twin model.

[0032] Preferably, the step of selecting a transfer learning model according to the source domain data, and inputting the source domain data into the transfer learning model for training to obtain a third digital twin model includes:

[0033] Acquire first data distribution information, first data type information, first data quality information, and first data label availability information according to the source domain data;

[0034] Acquire second data distribution information, second data type information, second data quality information, and second data label availability information according to the historical target data;

[0035] Determine a transfer learning model according to the first data distribution information, the first data type information, the first data quality information and the first data label availability information and the second data distribution information, the second data type information, the second data quality information and the second data label availability information;

[0036] Determining model parameters in the transfer learning model;

[0037] The source domain data is sequentially subjected to standardization and enhancement preprocessing to obtain preprocessed data;

[0038] The preprocessed data is input into the transfer learning model for training to obtain a third digital twin model.

[0039] Preferably, the step of optimizing the second digital twin model according to the transfer learning data to obtain the digital twin optimization model comprises:

[0040] Acquire historical target data, and plot the historical target data into a first histogram;

[0041] Acquire transfer learning data, and plot the transfer learning data into a second histogram;

[0042] Determining whether the overlap between the first histogram and the second histogram is greater than a preset value;

[0043] If the overlap between the first histogram and the second histogram is greater than a preset value, determining that the optimizer and loss function of the second digital twin model are the Adam optimizer and the cross entropy loss function respectively;

[0044] The second digital twin model is optimized according to the Adam optimizer and the cross entropy loss function to obtain the digital twin optimization model.

[0045] The present application also provides an aircraft engine digital twin modeling optimization system based on transfer learning, including:

[0046] A first acquisition module is used to acquire historical target data and multiple real-time target data of the aircraft engine, wherein the historical target data includes environmental characteristic data and control characteristic data;

[0047] A construction module, configured to use a random forest algorithm to construct a first digital twin model according to the environmental feature data and the control feature data;

[0048] A first training module, used for training the first digital twin model according to the plurality of real-time target data to obtain a second digital twin model;

[0049] A second acquisition module is used to acquire source domain data related to the historical target data;

[0050] A second training module is used to select a transfer learning model according to the source domain data, and input the source domain data into the transfer learning model for training to obtain a third digital twin model;

[0051] An optimization module is used to obtain transfer learning data between the third digital twin model and the second digital twin model, and optimize the second digital twin model according to the transfer learning data to obtain a digital twin optimization model.

[0052] Preferably, the building block comprises:

[0053] A first acquisition unit, configured to acquire atmospheric density data according to the environmental characteristic data, and acquire fuel flow data according to the control characteristic data;

[0054] A second acquisition unit is used to acquire the cross-sectional area of ​​the air inlet of the aircraft engine, the air temperature of the air inlet, the total pressure of the first airflow at the air inlet, and the total pressure of the second airflow at the exhaust port;

[0055] A construction unit is used to construct a first digital twin model according to the cross-sectional area, the intake temperature, the first airflow total pressure, the second airflow total pressure, the fuel flow data and the atmospheric density data, wherein the first digital twin model is:

[0056]

[0057] Among them, P(W) represents the predicted exhaust temperature, J(W) represents the intake temperature, R(L) represents the fuel flow data, J(M) represents the cross-sectional area, QY(1) represents the total pressure of the first airflow, QY(2) represents the total pressure of the second airflow, k represents the first empirical weight coefficient, n represents the second empirical weight coefficient, and m represents the third empirical weight coefficient.

[0058] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned aero-engine digital twin modeling optimization method based on transfer learning are implemented.

[0059] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned aero-engine digital twin modeling optimization method based on transfer learning are implemented.

[0060] The beneficial effects of the present invention are as follows: the present invention establishes a first digital twin model by combining environmental feature data and control feature data through a random forest algorithm, thereby being able to capture the performance and behavior changes of aircraft engines under different working conditions. Through transfer learning, the second digital twin model can learn some common features and patterns from the first digital twin model, and then perform fine-tuning and optimization based on real-time data. Such a method can reduce dependence on a large amount of real-time data and improve the generalization ability and adaptability of the model. The optimization of the second digital twin model can significantly improve the prediction accuracy and precision of the model. The third digital twin model can better adapt to possible data distribution changes and offsets in the target task through transfer learning. ,The transfer learning of source domain data helps to make the model more robust and able to effectively respond to these changes. Based on the transfer learning data of the third digital twin model, the parameters of the second digital twin model can be fine-tuned and optimized. This optimization can not only improve the prediction accuracy of the model, but also help improve the generalization ability of the model on new data. The optimized second digital twin model can more accurately monitor the operating status and health of aircraft engines in real time. The use of transfer learning data enables the model to adapt quickly when real-time data is updated and to detect potential problems or anomalies in time. By optimizing the second digital twin model through transfer learning, the risk of model prediction errors can be reduced, thereby improving the safety and reliability of aircraft engines. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 The figure is a schematic diagram of a method flow according to an embodiment of the present invention.

[0062] Figure 2 FIG. 1 is a schematic diagram of a device structure according to an embodiment of the present invention.

[0063] Figure 3 A schematic diagram of the internal structure of a computer device according to an embodiment of the present application.

[0064] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0065] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0066] like Figure 1-Figure 3 As shown, the present application provides an aircraft engine digital twin modeling optimization method based on transfer learning, including:

[0067] S1. Acquire historical target data and multiple real-time target data of an aircraft engine, wherein the historical target data includes environmental characteristic data and control characteristic data;

[0068] S2. Using a random forest algorithm to construct a first digital twin model according to the environmental feature data and the control feature data;

[0069] S3. Training the first digital twin model according to the plurality of real-time target data to obtain a second digital twin model;

[0070] S4, obtaining source domain data related to historical target data;

[0071] S5. Select a transfer learning model according to the source domain data, and input the source domain data into the transfer learning model for training to obtain a third digital twin model;

[0072] S6. Obtain transfer learning data between the third digital twin model and the second digital twin model, and optimize the second digital twin model according to the transfer learning data to obtain a digital twin optimization model.

[0073] As described in the above steps S1-S6, the present invention obtains historical target data and multiple real-time target data of the aircraft engine, wherein the historical target data includes environmental characteristic data and control characteristic data, and uses a random forest algorithm to construct a first digital twin model according to the environmental characteristic data and the control characteristic data, and trains the first digital twin model according to the multiple real-time target data to obtain a second digital twin model. The first digital twin model is established by combining the environmental characteristic data and the control characteristic data through the random forest algorithm. Since the first digital twin model is based on the historical target data, it can capture the performance and behavior changes of the aircraft engine under different working conditions, and provide a benchmark prediction model. In addition, the random forest algorithm is suitable for processing complex In the case of nonlinear relationships and a large number of features, it can effectively capture the potential patterns and laws in historical data. Since the second digital twin model is trained using multiple real-time target data, these real-time data may reflect the latest operating status of the current aircraft engine, environmental changes and other factors, and can update the model parameters in real time, thereby maintaining the accuracy and reliability of the model. Through transfer learning, the second digital twin model can learn some common features and patterns from the first digital twin model, and then fine-tune and optimize according to the real-time data. This method can reduce dependence on large amounts of real-time data and improve the generalization and adaptability of the model. The optimization of the second digital twin model can significantly improve the prediction accuracy and accuracy of the model, and it can be faster. Adapt to changes in engine working conditions, capture key features in real-time data, and thus more accurately predict engine performance and health status, obtain source domain data related to historical target data, select a transfer learning model based on the source domain data, and input the source domain data into the transfer learning model for training to obtain the third digital twin model. Since source domain data usually contains features and patterns that are related to but not completely the same as the target task, through transfer learning, the third digital twin model can use the knowledge and features in the source domain data to help the model better understand the target task data, which can improve the generalization ability of the model. Transfer learning can reduce the labeling and collection costs on the target domain data. By using existing source domain data, it can avoid the need for redundant data in the target task. A large amount of labeled data is required, which is particularly important for high-cost fields such as aircraft engines, because the acquisition and labeling of real-time data can be very expensive and difficult. Transfer learning based on source domain data can accelerate the model training process. Compared with training the model from scratch, transfer learning usually requires fewer iterations and tuning, allowing the model to be deployed to practical applications more quickly. The third digital twin model can better adapt to possible data distribution changes and offsets in the target task through transfer learning. Transfer learning of source domain data helps make the model more robust and able to effectively respond to these changes. The optimized third digital twin model can provide support for real-time decision-making. It can train more accurate and reliable models based on historical data and source domain data.Help monitor and predict the performance, health status and failure risk of aircraft engines, obtain transfer learning data between the third digital twin model and the second digital twin model, and optimize the second digital twin model according to the transfer learning data to obtain a digital twin optimization model. The third digital twin model can provide additional data features or feature adjustments, which may not be fully utilized or considered in the second digital twin model. Through transfer learning, these new features can be integrated into the second digital twin model, thereby improving its understanding and prediction capabilities of aircraft engine working conditions. Based on the transfer learning data of the third digital twin model, the parameters of the second digital twin model can be fine-tuned and optimized. This optimization can not only improve the prediction accuracy of the model, but also help improve the generalization ability of the model on new data. The optimized second digital twin model can more accurately monitor the operating status and health status of aircraft engines in real time. The use of transfer learning data enables the model to quickly adapt when real-time data is updated, and to detect potential problems or anomalies in a timely manner. By optimizing the second digital twin model through transfer learning, the risk of model prediction errors can be reduced, thereby improving the safety and reliability of aircraft engines. ,

[0074] In one embodiment, the step S2 of using a random forest algorithm to construct a first digital twin model according to the environmental feature data and the control feature data includes:

[0075] S21, acquiring atmospheric density data according to the environmental characteristic data, and acquiring fuel flow data according to the control characteristic data;

[0076] S22, obtaining a cross-sectional area of ​​an air inlet of the aircraft engine, an air temperature of the air inlet, a first airflow total pressure of the air inlet, and a second airflow total pressure of the exhaust port;

[0077] S23. Construct a first digital twin model according to the cross-sectional area, the intake temperature, the first airflow total pressure, the second airflow total pressure, the fuel flow data, and the atmospheric density data, wherein the first digital twin model is:

[0078]

[0079] Among them, P(W) represents the predicted exhaust temperature, J(W) represents the intake temperature, R(L) represents the fuel flow data, J(M) represents the cross-sectional area, QY(1) represents the total pressure of the first airflow, QY(2) represents the total pressure of the second airflow, k represents the first empirical weight coefficient, n represents the second empirical weight coefficient, and m represents the third empirical weight coefficient.

[0080] As described in the above steps S21-S23, the present invention obtains atmospheric density data through environmental characteristic data, obtains fuel flow data according to control characteristic data, obtains the cross-sectional area of ​​the aircraft engine air inlet, the intake temperature of the air inlet, the first airflow total pressure of the air inlet and the second airflow total pressure of the exhaust port, and constructs a first digital twin model based on the cross-sectional area, intake temperature, first airflow total pressure, second airflow total pressure, fuel flow data and atmospheric density data. By including data such as atmospheric density, fuel flow, physical parameters of the air inlet and exhaust port, the first digital twin model can more comprehensively simulate the working state of the aircraft engine. These parameters directly affect the combustion process, intake and exhaust performance of the engine, and are important factors in evaluating engine performance and health status. The model constructed using these physical parameters can more accurately capture the complex physical and chemical processes inside the engine. For example, atmospheric density affects the air-fuel mixture ratio, while fuel flow directly affects the combustion rate and engine output power. Taking into account environmental characteristic data such as atmospheric density, the model can better adapt to different environmental conditions, such as high altitude areas or situations with large temperature changes, which helps to improve the model's prediction accuracy and reliability in practical applications. The first digital twin model can reflect the current working status and performance of the aircraft engine in real time by combining real-time fuel flow, total airflow pressure and other data. The first digital twin model constructed using environmental characteristic data and control characteristic data can better adapt to new working conditions and data distribution in target tasks. This enhanced adaptability helps the application of the model in different flight missions or different engine configurations.

[0081] In one embodiment, the step S21 of acquiring the atmospheric density data according to the environmental characteristic data and acquiring the fuel flow rate data according to the control characteristic data comprises:

[0082] S211, acquiring temperature parameters, pressure parameters and altitude parameters according to the environmental characteristic data;

[0083] S212, obtaining the temperature gradient and gas constant of the environment in which the aircraft engine is located;

[0084] S213, calculating atmospheric density data according to the temperature parameter, pressure parameter, altitude parameter, temperature gradient and gas constant, wherein the calculation formula is:

[0085]

[0086] Among them, Q(M) represents atmospheric density data, Q(C) represents pressure parameters, W(C) represents temperature parameters, Q(C) represents gas constant, W(T) represents temperature gradient, and H(C) represents altitude parameters;

[0087] S214, acquiring a throttle lever angle parameter, a flight speed parameter, and a throttle opening parameter according to the control characteristic data;

[0088] S215, obtaining a basic fuel flow rate of the aircraft engine;

[0089] S216, calculating fuel flow data according to the basic fuel flow, throttle lever angle parameter, flight speed parameter and throttle opening parameter, wherein the calculation formula is:

[0090] R(L)=J(R)*[a*Y(J)+b*F(S)+c*Y(K)];

[0091] Wherein, R(L) represents the fuel flow data, J(R) represents the basic fuel flow, a represents the weight of the throttle lever angle parameter, Y(J) represents the throttle lever angle parameter, b represents the weight of the flight speed parameter, F(S) represents the flight speed parameter, c represents the weight of the throttle opening parameter, and Y(K) represents the throttle opening parameter.

[0092] As described in the above steps S211-S216, the present invention obtains temperature parameters, pressure parameters and altitude parameters through environmental characteristic data, as well as the temperature gradient and gas constant in the environment where the aircraft engine is located, and calculates atmospheric density data based on the temperature parameters, pressure parameters, altitude parameters, temperature gradient and gas constant. By considering environmental parameters such as temperature, pressure, altitude, as well as temperature gradient and gas constant, the constructed atmospheric density model can more accurately reflect the physical state of the environment in which the aircraft engine is located. This accuracy is crucial for the prediction and simulation of the model, and can improve the accuracy and practicality of the model. Taking into account different temperature, pressure and altitude parameters, the model can adapt to various working conditions and environmental scenarios. This adaptability makes the model more robust and reliable when facing new data or changing working conditions. Accurately calculated atmospheric density data can help optimize the model's ability to predict aircraft engine performance and combustion processes. Atmospheric density directly affects the air-fuel ratio and combustion efficiency, and is very important for evaluating aircraft engine output. Output power and fuel efficiency are crucial. By comprehensively considering multiple environmental parameters and physical characteristics, especially the accurate calculation of atmospheric density, the prediction error of the model can be effectively reduced, and the prediction ability, adaptability and real-time monitoring support ability of the model can be significantly improved, thereby effectively optimizing the performance management and health monitoring of aircraft engines. By controlling the characteristic data, the throttle lever angle parameters, flight speed parameters and throttle opening parameters are obtained, and the basic fuel flow of the aircraft engine is obtained. The fuel flow data is calculated based on the basic fuel flow, throttle lever angle parameters, flight speed parameters and throttle opening parameters. By combining multiple key parameters such as throttle lever angle, flight speed, throttle opening and basic fuel flow, the model can more comprehensively understand and simulate the complex process of engine fuel consumption. This comprehensive analysis can improve the accuracy and precision of fuel flow prediction. By accurately calculating and predicting fuel flow, it can help optimize the fuel consumption and operating efficiency of the engine. Based on accurate fuel flow data, the model can monitor the performance status of the engine in real time.

[0093] In one embodiment, the step S3 of training the first digital twin model according to the plurality of real-time target data to obtain the second digital twin model includes:

[0094] S31, select any real-time target data as a training set and input it into the first digital twin model for training, and input the remaining real-time target data as a verification set into the first digital twin model for verification, to obtain multiple training results;

[0095] S32. Adjust the model parameters of the first digital twin model according to the multiple training results to obtain a second digital twin model.

[0096] As described in the above steps S31-S32, the present invention selects different real-time target data as training sets and verification sets, so that the first digital twin model can be trained and verified under a variety of different data distributions. This diversified training helps the model to better adapt to different working conditions and operating scenarios, improve its generalization ability and robustness, and adjust the model parameters of the first digital twin model based on multiple training results. The structure and parameter settings of the model can be optimized by comparing different training effects. This method can help find a model configuration that is more suitable for the current task and data characteristics, improve the performance and effect of the model, and the optimized second digital twin model can process real-time data more quickly and make accurate predictions or decisions, which is crucial for real-time monitoring, health management and performance optimization of aircraft engines. By effectively utilizing training data and optimizing model structure, the cost of model training and deployment can be reduced, which not only saves time and resources, but also improves the actual application efficiency and cost-effectiveness of the model. By comparing and verifying multiple training results, the interpretability and trustworthiness of the model can be enhanced.

[0097] In one embodiment, the step S5 of selecting a transfer learning model according to the source domain data and inputting the source domain data into the transfer learning model for training to obtain a third digital twin model includes:

[0098] S51, acquiring first data distribution information, first data type information, first data quality information, and first data label availability information according to the source domain data;

[0099] S52, acquiring second data distribution information, second data type information, second data quality information, and second data label availability information according to the historical target data;

[0100] S53, determining a transfer learning model according to the first data distribution information, the first data type information, the first data quality information and the first data label availability information and the second data distribution information, the second data type information, the second data quality information and the second data label availability information;

[0101] S54, determining model parameters in the transfer learning model;

[0102] S55, performing standardization and enhancement preprocessing on the source domain data in sequence to obtain preprocessed data;

[0103] S56. Input the preprocessed data into the transfer learning model for training to obtain a third digital twin model.

[0104] As described in the above steps S51-S56, the present invention obtains first data distribution information, first data type information, first data quality information and first data label availability information through source domain data and obtains second data distribution information, second data type information, second data quality information and second data label availability information through historical target data, and determines a transfer learning model based on the first data distribution information, first data type information, first data quality information and first data label availability information and the second data distribution information, second data type information, second data quality information and second data label availability information. By comparing the distribution information of the source domain data and the historical target data, their similarities or differences in the feature space can be evaluated, which helps to determine When can the source domain data be directly applied, and when is adaptive adjustment or specific data preprocessing required to ensure data consistency and availability? Analyzing the type (such as numeric type, text type, etc.) and quality (such as data integrity, missing values, etc.) of the first and second data helps to understand the applicability and processing requirements of the data. For example, if the quality of historical target data is low, additional data cleaning or supplementation may be required to improve the stability and accuracy of the model. Evaluating the label availability of the first and second data, that is, whether the data has been labeled and what the quality of the labeling is, is particularly important for supervised learning models. This can guide the model to select appropriate learning methods and training strategies to maximize the use of available labeled data. Based on the distribution, type, quality, and labeling of the data, the model can select the appropriate learning method and training strategy. Based on the availability information of the signature, you can choose the appropriate transfer learning method and strategy. For example, if the source domain data and the target data are similar in data distribution, you can directly apply simple transfer learning technology. If the difference is large, a more complex domain adaptation method is required to adjust the model. Based on comprehensive data analysis and evaluation, the constructed transfer learning model can provide more reliable and accurate prediction results, which is of great significance for improving engine performance, reducing maintenance costs and improving flight safety. By determining the model parameters in the transfer learning model, the source domain data is standardized and enhanced preprocessed in turn to obtain preprocessed data, and then the preprocessed data is input into the transfer learning model for training to obtain the third digital twin model. Through standardization and enhanced preprocessing, it can be improved. The consistency and quality of the source domain data ensure that the data has similar data distribution and feature representation before entering the model, which helps reduce the risk of model training difficulties or performance degradation caused by data inconsistency and enhances the stability and reliability of the model. By training the third digital twin model through the transfer learning model, the model can be further adapted to new target tasks and data environments based on the source domain data and historical target data. This adaptability enables the model to be better generalized to different operating conditions and environments, improving the practicality and adaptability of the model. By determining appropriate transfer learning model parameters, the performance and efficiency of the model can be optimized to better meet the needs of digital twin modeling of aircraft engines. The optimized data is standardized and enhanced preprocessed.It can improve the efficiency and speed of model training. The third digital twin model can provide more accurate and reliable prediction results and decision support after optimized training and preprocessing, which is of great significance for aircraft engine performance optimization, fault prediction and safety management, and helps to improve the efficiency and safety of aviation operations.

[0105] In one embodiment, the step S6 of optimizing the second digital twin model according to the transfer learning data to obtain the digital twin optimization model includes:

[0106] S61, acquiring historical target data, and plotting the historical target data into a first histogram;

[0107] S62, acquiring transfer learning data, and plotting the transfer learning data into a second histogram;

[0108] S63, determining whether the overlap between the first histogram and the second histogram is greater than a preset value;

[0109] If the overlap between the first histogram and the second histogram is greater than a preset value, determining that the optimizer and loss function of the second digital twin model are the Adam optimizer and the cross entropy loss function respectively;

[0110] S64. Optimize the second digital twin model according to the Adam optimizer and the cross entropy loss function to obtain a digital twin optimization model.

[0111] As described in the above steps S101-S103, the present invention obtains historical target data and transfer learning data, and plots the historical target data and the transfer learning data into a first histogram and a second histogram respectively, and determines whether the degree of overlap between the first histogram and the second histogram is greater than a preset value. If so, it is determined that the optimizer and loss function of the second digital twin model are the Adam optimizer and the cross entropy loss function respectively, and then the second digital twin model is optimized according to the Adam optimizer and the cross entropy loss function to obtain a digital twin optimization model. By plotting the histograms of the historical target data and the transfer learning data and comparing their overlap, it is possible to intuitively evaluate whether the distribution of the data is similar. If the degree of overlap of the histograms of the two is higher than the preset value, it indicates that the data is relatively consistent in the feature space and is suitable for direct application in the transfer learning model, which helps to ensure the generalization ability and stability of the model on new data. If it is determined that the degree of overlap of the histograms meets the preset conditions, the Adam optimizer can be selected. The Adam optimizer can usually effectively adjust the learning rate and adaptively adjust the gradients of different features, which is suitable for processing variable aero-engine data. The cross-entropy loss function performs well in classification tasks and can help the model effectively learn and optimize the prediction results, especially for tasks such as anomaly detection and fault prediction. It has high sensitivity and accuracy. Using suitable optimizers and loss functions to optimize the second digital twin model can improve the performance and prediction accuracy of the model in aero-engine digital twin modeling, which is crucial for real-time monitoring of engine status, predicting faults, and optimizing maintenance plans, and helps to improve aviation safety and operational efficiency. By making full use of the histogram analysis of historical target data and transfer learning data, additional data processing costs and model adjustment time can be reduced. This method not only improves the efficiency of model construction, but also can quickly respond to different data environments and task requirements, reducing technical risks and costs in the modeling process.

[0112] The present application also provides an aircraft engine digital twin modeling optimization system based on transfer learning, including:

[0113] A first acquisition module is used to acquire historical target data and multiple real-time target data of the aircraft engine, wherein the historical target data includes environmental characteristic data and control characteristic data;

[0114] A construction module, configured to use a random forest algorithm to construct a first digital twin model according to the environmental feature data and the control feature data;

[0115] A first training module, used for training the first digital twin model according to the plurality of real-time target data to obtain a second digital twin model;

[0116] A second acquisition module is used to acquire source domain data related to the historical target data;

[0117] A second training module is used to select a transfer learning model according to the source domain data, and input the source domain data into the transfer learning model for training to obtain a third digital twin model;

[0118] An optimization module is used to obtain transfer learning data between the third digital twin model and the second digital twin model, and optimize the second digital twin model according to the transfer learning data to obtain a digital twin optimization model.

[0119] In one embodiment, the building block comprises:

[0120] A first acquisition unit, configured to acquire atmospheric density data according to the environmental characteristic data, and acquire fuel flow data according to the control characteristic data;

[0121] A second acquisition unit is used to acquire the cross-sectional area of ​​the air inlet of the aircraft engine, the air temperature of the air inlet, the total pressure of the first airflow at the air inlet, and the total pressure of the second airflow at the exhaust port;

[0122] A construction unit is used to construct a first digital twin model according to the cross-sectional area, the intake temperature, the first airflow total pressure, the second airflow total pressure, the fuel flow data and the atmospheric density data, wherein the first digital twin model is:

[0123]

[0124] Among them, P(W) represents the predicted exhaust temperature, J(W) represents the intake temperature, R(L) represents the fuel flow data, J(M) represents the cross-sectional area, QY(1) represents the total pressure of the first airflow, QY(2) represents the total pressure of the second airflow, k represents the first empirical weight coefficient, n represents the second empirical weight coefficient, and m represents the third empirical weight coefficient.

[0125] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned aero-engine digital twin modeling optimization method based on transfer learning are implemented.

[0126] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned aero-engine digital twin modeling optimization method based on transfer learning are implemented.

[0127] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0128] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0129] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for optimizing digital twin modeling of an aero-engine based on transfer learning, characterized in that: include: Acquire historical target data and multiple real-time target data of the aircraft engine, wherein the historical target data includes environmental characteristic data and control characteristic data; Using a random forest algorithm to construct a first digital twin model according to the environmental characteristic data and the control characteristic data; Training the first digital twin model according to the plurality of real-time target data to obtain a second digital twin model; Obtain source domain data related to historical target data; Selecting a transfer learning model according to the source domain data, and inputting the source domain data into the transfer learning model for training to obtain a third digital twin model; Acquire transfer learning data between the third digital twin model and the second digital twin model, and optimize the second digital twin model according to the transfer learning data to obtain a digital twin optimization model.

2. The aero-engine digital twin modeling optimization method based on transfer learning according to claim 1 is characterized in that: The step of using a random forest algorithm to construct a first digital twin model according to the environmental feature data and the control feature data includes: Acquire atmospheric density data according to the environmental characteristic data, and acquire fuel flow data according to the control characteristic data; Obtaining a cross-sectional area of ​​an air inlet of an aircraft engine, an air temperature of the air inlet, a first airflow total pressure of the air inlet, and a second airflow total pressure of the exhaust port; A first digital twin model is constructed according to the cross-sectional area, the intake temperature, the first airflow total pressure, the second airflow total pressure, the fuel flow data, and the atmospheric density data, wherein the first digital twin model is: Among them, P(W) represents the predicted exhaust temperature, J(W) represents the intake temperature, R(L) represents the fuel flow data, J(M) represents the cross-sectional area, QY(1) represents the total pressure of the first airflow, QY(2) represents the total pressure of the second airflow, k represents the first empirical weight coefficient, n represents the second empirical weight coefficient, and m represents the third empirical weight coefficient.

3. The aero-engine digital twin modeling optimization method based on transfer learning according to claim 2 is characterized in that: The step of acquiring the atmospheric density data according to the environmental characteristic data and acquiring the fuel flow rate data according to the control characteristic data comprises: Acquire temperature parameters, pressure parameters and altitude parameters according to the environmental characteristic data; Obtain the temperature gradient and gas constant of the environment in which the aircraft engine is located; The atmospheric density data is calculated according to the temperature parameter, pressure parameter, altitude parameter, temperature gradient and gas constant, wherein the calculation formula is: Among them, Q(M) represents atmospheric density data, Y(C) represents pressure parameters, W(C) represents temperature parameters, Q(C) represents gas constant, W(T) represents temperature gradient, and H(C) represents altitude parameters; Acquire a throttle lever angle parameter, a flight speed parameter, and a throttle opening parameter according to the control characteristic data; Obtain basic fuel flow for aircraft engines; The fuel flow data is calculated according to the basic fuel flow, the throttle lever angle parameter, the flight speed parameter and the throttle opening parameter, wherein the calculation formula is: R(L)=J(R)*[a*Y(J)+b*F(S)+c*Y(K)]; Among them, R(L) represents the fuel flow data, H(R) represents the basic fuel flow, a represents the weight of the throttle lever angle parameter, Y(J) represents the throttle lever angle parameter, b represents the weight of the flight speed parameter, F(S) represents the flight speed parameter, c represents the weight of the throttle opening parameter, and Y(k) represents the throttle opening parameter.

4. The aero-engine digital twin modeling optimization method based on transfer learning according to claim 1 is characterized in that: The step of training the first digital twin model according to the plurality of real-time target data to obtain the second digital twin model comprises: Select any real-time target data as a training set and input it into the first digital twin model for training, and input the remaining real-time target data as a verification set into the first digital twin model for verification, to obtain multiple training results; The model parameters of the first digital twin model are adjusted according to the multiple training results to obtain the second digital twin model.

5. The aero-engine digital twin modeling optimization method based on transfer learning according to claim 1 is characterized in that: The step of selecting a transfer learning model according to the source domain data, and inputting the source domain data into the transfer learning model for training to obtain a third digital twin model includes: Acquire first data distribution information, first data type information, first data quality information, and first data label availability information according to the source domain data; Acquire second data distribution information, second data type information, second data quality information, and second data label availability information according to the historical target data; Determine a transfer learning model according to the first data distribution information, the first data type information, the first data quality information and the first data label availability information and the second data distribution information, the second data type information, the second data quality information and the second data label availability information; Determining model parameters in the transfer learning model; The source domain data is sequentially subjected to standardization and enhancement preprocessing to obtain preprocessed data; The preprocessed data is input into the transfer learning model for training to obtain a third digital twin model.

6. The aero-engine digital twin modeling optimization method based on transfer learning according to claim 1 is characterized in that: The step of optimizing the second digital twin model according to the transfer learning data to obtain a digital twin optimization model includes: Acquire historical target data, and plot the historical target data into a first histogram; Acquire transfer learning data, and plot the transfer learning data into a second histogram; Determining whether the overlap between the first histogram and the second histogram is greater than a preset value; If the overlap between the first histogram and the second histogram is greater than a preset value, determining that the optimizer and loss function of the second digital twin model are the Adam optimizer and the cross entropy loss function respectively; The second digital twin model is optimized according to the Adam optimizer and the cross entropy loss function to obtain the digital twin optimization model.

7. An aircraft engine digital twin modeling optimization system based on transfer learning, characterized in that: include: A first acquisition module is used to acquire historical target data and multiple real-time target data of the aircraft engine, wherein the historical target data includes environmental characteristic data and control characteristic data; A construction module, configured to use a random forest algorithm to construct a first digital twin model according to the environmental feature data and the control feature data; A first training module, used for training the first digital twin model according to the plurality of real-time target data to obtain a second digital twin model; A second acquisition module is used to acquire source domain data related to the historical target data; A second training module is used to select a transfer learning model according to the source domain data, and input the source domain data into the transfer learning model for training to obtain a third digital twin model; An optimization module is used to obtain transfer learning data between the third digital twin model and the second digital twin model, and optimize the second digital twin model according to the transfer learning data to obtain a digital twin optimization model.

8. The aero-engine digital twin modeling and optimization system based on transfer learning according to claim 7 is characterized in that: The building blocks include: A first acquisition unit, configured to acquire atmospheric density data according to the environmental characteristic data, and to acquire fuel flow data according to the control characteristic data; A second acquisition unit is used to acquire the cross-sectional area of ​​the air inlet of the aircraft engine, the air temperature of the air inlet, the total pressure of the first airflow at the air inlet, and the total pressure of the second airflow at the exhaust port; A construction unit is used to construct a first digital twin model according to the cross-sectional area, the intake temperature, the first airflow total pressure, the second airflow total pressure, the fuel flow data and the atmospheric density data, wherein the first digital twin model is: Among them, P(W) represents the predicted exhaust temperature, J(W) represents the intake temperature, R(L) represents the fuel flow data, J(M) represents the cross-sectional area, QY(1) represents the total pressure of the first airflow, QY(2) represents the total pressure of the second airflow, k represents the first empirical weight coefficient, n represents the second empirical weight coefficient, and m represents the third empirical weight coefficient.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images