Cascade modeling method and device for aero-engine multi-task learning

Abstract

The application relates to a kind of aero-engine multi-task learning cascade modeling method and device.Method comprising: obtaining the detection data of each actual measuring point of the structure of each component of aero-engine, the task association relationship between each component structure and the gas path flow direction information, and the flight condition information of aero-engine, and generating the engine task corresponding to each component structure; identify the task dependency information between each component structure, and construct low-fidelity network and high-fidelity network, and based on low-fidelity network, high-fidelity network and the gas path flow direction information between each component structure, the engine performance model of aero-engine is constructed.The simulation effect of the engine performance model of aero-engine is improved by using the method.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of simulation modeling and multi-task learning, in particular to a cascade modeling method and device for aero-engine multi-task learning. BACKGROUND

[0002] The modeling technology of an aero-engine is the basis for tasks such as whole-machine control, fault diagnosis, and fault-tolerant control, and has been widely concerned by practitioners in the field of aero-engines for a long time. However, due to the multidisciplinary coupling of the engine and the complexity of the system, it is extremely difficult to establish an engine performance model covering the whole process of the engine. Therefore, how to establish an engine performance model covering the whole process of the engine is the current research focus.

[0003] The traditional modeling method of the engine performance model is to use a data-driven method as a residual compensation model between a physical model and actual operation data. However, this kind of method faces the problem of insufficient generalization of the data-driven model, and needs to rely on large-scale real data, thereby leading to poor simulation effect of the engine performance model of the aero-engine. SUMMARY

[0004] Therefore, it is necessary to provide a cascade modeling method and device for aero-engine multi-task learning, a computer device, a computer readable storage medium, and a computer program product in view of the above technical problems.

[0005] In a first aspect, the present application provides a cascade modeling method for aero-engine multi-task learning, comprising:

[0006] obtaining detection data of each actual measuring point of each component structure of an aero-engine, a task association relationship between each component structure, air path flow direction information between each component structure, and flight condition information of the aero-engine, and generating an engine task corresponding to each component structure based on the detection data of each actual measuring point of each component structure and the air path flow direction information between each component structure;

[0007] identifying task dependency information between each component structure based on the task association relationship between each component structure and the engine task corresponding to each component structure, and constructing a low-fidelity network based on the flight condition information and the engine task corresponding to each component structure;

[0008] constructing a high-fidelity network based on the low-fidelity network, the flight condition information, and the task dependency information between each component structure, and constructing an engine performance model of the aero-engine based on the low-fidelity network, the high-fidelity network, and the air path flow direction information between each component structure.

[0009] Optionally, the detection data of each actual measuring point of each component structure and the air path flow direction information between each component structure are used to generate the engine task corresponding to each component structure, including:

[0010] The detection data of each actual measuring point of each component structure is used to identify the detection type of each actual measuring point.

[0011] The detection type of each actual measuring point is used to query the operation task type corresponding to each component structure in a component operation database.

[0012] The operation task type corresponding to each component structure and the air path flow direction information between each component structure are used to query the engine task of each component structure in an engine operation process in an engine operation database.

[0013] Optionally, the task dependency information between each component structure is identified based on the task association relationship between each component structure and the engine task corresponding to each component structure, including:

[0014] The task association relationship between each component structure is used to identify the information transmission mode between each component structure.

[0015] The input demand information corresponding to each component structure and the output information corresponding to each component structure are identified based on the engine task corresponding to each component structure, and the data interaction information between each component structure is identified based on the information transmission mode between each component structure, the input demand information corresponding to each component structure, and the output information corresponding to each component structure.

[0016] The data interaction information between each component structure is used as the task dependency information between each component structure.

[0017] Optionally, the low-fidelity network is constructed based on the flight condition information and the engine task corresponding to each component structure, including:

[0018] Baseline simulation data of the aero-engine is obtained.

[0019] Each flight condition data in the flight condition information is identified, and an initial low-fidelity network is trained based on each baseline simulation data and each flight condition data to obtain a low-fidelity network.

[0020] Optionally, the high-fidelity network is constructed based on the low-fidelity network, the flight condition information, and the task dependency information between each component structure, including:

[0021] Based on the data interaction information between the structures of the components, sub-data interaction information between the detection data of the actual measuring points is identified, and based on the flight condition data, model parameters are obtained through the low-fidelity network;

[0022] Based on the detection data of the actual measuring points, the flight condition data, and the model parameters, an initial high-fidelity network is trained to obtain a high-fidelity network.

[0023] Optionally, based on the low-fidelity network, the high-fidelity network, and the gas path flow direction information between the structures of the components, an engine performance model of the aero-engine is constructed, including:

[0024] Physical information of each component structure of the aero-engine is obtained, and based on the physical information of each component structure, a physical model of the aero-engine is constructed;

[0025] Based on the low-fidelity network and the high-fidelity network, a data-driven model of the aero-engine is constructed through a multi-task cascade modeling strategy;

[0026] The physical model and the data-driven model are used as the engine performance model of the aero-engine.

[0027] In a second aspect, the application further provides a cascade modeling device for multi-task learning of an aero-engine, including:

[0028] An acquisition module is configured to acquire detection data of actual measuring points of each component structure of an aero-engine, a task association relationship between the component structures, gas path flow direction information between the component structures, and flight condition information of the aero-engine, and based on the detection data of the actual measuring points of each component structure and the gas path flow direction information between the component structures, generate corresponding engine tasks of the component structures.

[0029] An identification module is configured to identify task dependency information between the component structures based on the task association relationship between the component structures and the corresponding engine tasks of the component structures, and based on the flight condition information and the corresponding engine tasks of the component structures, construct a low-fidelity network.

[0030] A construction module is configured to construct a high-fidelity network based on the low-fidelity network, the flight condition information, and the task dependency information between the component structures, and based on the low-fidelity network, the high-fidelity network, and the gas path flow direction information between the component structures, construct an engine performance model of the aero-engine.

[0031] Optionally, the acquisition module is specifically configured to:

[0032] identify a detection type of each actual measurement point based on the detection data of each actual measurement point of each component structure;

[0033] query a corresponding operation task type of each component structure in the component operation database based on the detection type of each actual measurement point;

[0034] query an engine task of each component structure in the engine operation database based on the corresponding operation task type of each component structure and the gas path flow direction information between the component structures.

[0035] Optionally, the identification module is specifically configured to:

[0036] identify an information transmission mode between the component structures based on the task association relationship between the component structures;

[0037] identify input demand information corresponding to each component structure and output information corresponding to each component structure based on the engine task corresponding to each component structure, and identify data interaction information between the component structures based on the information transmission mode between the component structures, the input demand information corresponding to each component structure, and the output information corresponding to each component structure;

[0038] take the data interaction information between the component structures as task dependency information between the component structures.

[0039] Optionally, the identification module is specifically configured to:

[0040] acquire baseline simulation data of the aero-engine;

[0041] identify each flight condition data in the flight condition information, and train an initial low-fidelity network based on the baseline simulation data and the flight condition data to obtain a low-fidelity network.

[0042] Optionally, the construction module is specifically configured to:

[0043] identify sub-data interaction information between the detection data of each actual measurement point based on the data interaction information between the component structures, and acquire each model parameter through the low-fidelity network based on the flight condition data;

[0044] train an initial high-fidelity network based on the detection data of each actual measurement point, the flight condition data, and the model parameter to obtain a high-fidelity network.

[0045] Optionally, the construction module is specifically used for:

[0046] acquiring physical information of each component structure of the aero-engine, and constructing a physical model of the aero-engine based on the physical information of each component structure;

[0047] constructing a data-driven model of the aero-engine by a multi-task cascade modeling strategy based on the low-fidelity network and the high-fidelity network;

[0048] using the physical model and the data-driven model as an engine performance model of the aero-engine.

[0049] In a third aspect, the present application provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method in any one of the first aspect when executing the computer program.

[0050] In a fourth aspect, the present application provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program implements the steps of the method in any one of the first aspect when executed by a processor.

[0051] In a fifth aspect, the present application provides a computer program product. The computer program product comprises a computer program, and the computer program implements the steps of the method in any one of the first aspect when executed by a processor.

[0052] The cascade modeling method and device for aero-engine multi-task learning have the following advantages. The detection data of each actual measuring point of each component structure of an aero-engine, the task association relationship between each component structure, the gas path flow direction information between each component structure, and the flight condition information of the aero-engine are obtained, and the engine tasks corresponding to each component structure are generated based on the detection data of each actual measuring point of each component structure and the gas path flow direction information between each component structure. The task dependency information between each component structure is identified based on the task association relationship between each component structure and the engine tasks corresponding to each component structure, and a low-fidelity network is constructed based on the flight condition information and the engine tasks corresponding to each component structure. A high-fidelity network is constructed based on the low-fidelity network, the flight condition information, and the task dependency information between each component structure, and an engine performance model of the aero-engine is constructed based on the low-fidelity network, the high-fidelity network, and the gas path flow direction information between each component structure. In this scheme, the combination of low-fidelity and high-fidelity data is adopted, so that the precision and robustness of the model can be effectively improved without significantly increasing the cost, and the method is suitable for resource-limited operating environments. Moreover, the physical information neural network model of multi-task cascade is adopted in this scheme, so that the fitting ability of the data-driven method is improved while the interpretability of the physical model is retained, and the method is suitable for complex engine system modeling. Finally, the modeling method integrates the physical model and the data-driven model, overcomes the error problem between the traditional model and the actual measurement, and has higher prediction precision and correction efficiency. Therefore, the simulation effect of the engine performance model of the aero-engine is comprehensively improved. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments or the related art, the drawings needed to be used in the embodiments or the related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0054] Figure 1 A flowchart of an aero-engine multi-task learning cascade modeling method in an embodiment;

[0055] Figure 2 A model structure diagram of a data-driven model in an embodiment;

[0056] Figure 3 A schematic diagram of a first set of flight condition curves in an embodiment;

[0057] Figure 4A schematic diagram of simulation results of PBM and ROE measurement points under a first set of flight conditions in an embodiment;

[0058] Figure 5 A schematic diagram of a schematic curve under a second set of flight conditions in an embodiment;

[0059] Figure 6 A schematic diagram of simulation results after correction based on MFPINN in an embodiment;

[0060] Figure 7 A flowchart of an example of cascade modeling of aero-engine multi-task learning in an embodiment;

[0061] Figure 8 A block diagram of a cascade modeling device for aero-engine multi-task learning in an embodiment;

[0062] Figure 9 An internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0063] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0064] The cascade modeling method for aero-engine multi-task learning provided by the embodiments of the present application can be applied to the application environment of cascade modeling of aero-engine multi-task learning. The terminal can be, but is not limited to, various personal computers, notebook computers, medium-sized computers, etc. The terminal uses a combination of low-fidelity and high-fidelity data, so that the method can effectively improve the precision and robustness of the model without significantly increasing the cost, and is suitable for resource-limited operating environments. Moreover, the physical information neural network model through multi-task cascade makes the method not only retain the interpretability of the physical model, but also improve the fitting ability of the data-driven method, which is suitable for complex engine system modeling. Finally, the modeling method integrates the physical model and the data-driven model, overcomes the error problem between the traditional model and the actual measurement, and has higher prediction accuracy and correction efficiency. Thus, the simulation effect of constructing the engine performance model of the aero-engine is comprehensively improved.

[0065] In an exemplary embodiment, as shown in Figure 1 A cascade modeling method for aero-engine multi-task learning is provided, which is taken as an example of the method applied to a terminal and includes the following steps S101-S103. Among them:

[0066] In step S101, detection data of each actual measuring point of each component structure of the aero-engine, task association relationship between each component structure, air path flow direction information between each component structure, and flight condition information of the aero-engine are acquired, and based on the detection data of each actual measuring point of each component structure and the air path flow direction information between each component structure, an engine task corresponding to each component structure is generated.

[0067] In this embodiment, the terminal acquires detection data of each actual measuring point of each component structure of the aero-engine. The actual measuring point is arranged on each component structure of the aero-engine and used to identify a component state of each component and an actual engine task corresponding to the engine task of the component structure. Each component structure includes but is not limited to a component structure such as a compressor, a combustion chamber, a turbine, and an exhaust device. The actual measuring point corresponding to each component structure includes, for example, an outlet static pressure and a rotating speed of the compressor, and an outlet temperature of the turbine. Through the flow direction of the air flow, the task of the compressor is defined as an upstream task, and the task of the turbine is defined as a downstream task. The establishment of such task relationship reflects the causal dependence in the physical system. Then, the terminal acquires a task association relationship between each component structure, which includes a task execution sequence and an information transmission mode between each component. Then, the terminal acquires air path flow direction information between each component structure and flight condition information of the aero-engine. The flight condition information includes flight condition data of multiple different flight condition types, such as a flight power type, a flight time type, and a flight power consumption type. Finally, based on the detection data of each actual measuring point of each component structure and the air path flow direction information between each component structure, an engine task corresponding to each component structure is generated. The specific generation process will be described in detail later.

[0068] In step S102, based on the task association relationship between each component structure and the engine task corresponding to each component structure, task dependency information between each component structure is identified, and based on the flight condition information and the engine task corresponding to each component structure, a low-fidelity network is constructed.

[0069] In this embodiment, the terminal identifies the task dependency information between the component structures based on the task association relationship between the component structures and the engine tasks corresponding to the component structures, and constructs a low-fidelity network based on the flight condition information and the engine tasks corresponding to the component structures. The low-fidelity network is constructed by a Multi-Fidelity Physics-Informed Neural Network (MFPINN) for rapid correction of the engine model. The task dependency information between the component structures is the data interaction information between the detection data corresponding to each actual measurement point between the component structures, that is, the data interaction information between the component structures. The specific identification process and the construction process will be described in detail later. The low-fidelity network is used to generate model parameters of a data-driven model approximating the baseline condition.

[0070] In step S103, a high-fidelity network is constructed based on the low-fidelity network, the flight condition information, and the task dependency information between the component structures, and an engine performance model of the aero-engine is constructed based on the low-fidelity network, the high-fidelity network, and the air path flow direction information between the component structures.

[0071] In this embodiment, the terminal constructs a high-fidelity network based on the low-fidelity network, the flight condition information, and the task dependency information between the component structures, and constructs an engine performance model of the aero-engine based on the low-fidelity network, the high-fidelity network, and the air path flow direction information between the component structures. The high-fidelity network is constructed by a Multi-Fidelity Physics-Informed Neural Network (MFPINN) for rapid correction of the engine model. The engine performance model includes a physical model and a data-driven model, and the formula of the data-driven model is as follows:

[0072]

[0073] wherein, NN L is the low-fidelity network, NN Hnl , and NN Hn is the high-fidelity network. is the flight condition data of each flight condition type. As shown in Figure 2 , it is a model structure diagram of the data-driven model.

[0074] Based on the above scheme, by adopting the combination of low-fidelity and high-fidelity data, the method can effectively improve the precision and robustness of the model without significantly increasing the cost, and is suitable for resource-limited operating environment. And the scheme through the multi-task cascaded physical information neural network model, the method not only retains the explainability of the physical model, but also improves the fitting ability of the data-driven method, which is suitable for complex engine system modeling. Finally, the modeling method integrates the physical model and the data-driven model, overcomes the error problem between the traditional model and the actual measurement, and has higher prediction accuracy and correction efficiency. Thus, the simulation effect of constructing the engine performance model of the aero-engine is comprehensively improved.

[0075] Optionally, based on the detection data of each actual measuring point of each component structure and the gas path flow direction information between the component structures, an engine task corresponding to each component structure is generated, including: identifying the detection type of each actual measuring point based on the detection data of each actual measuring point of each component structure; querying the running task type corresponding to each component structure in the component running database based on the detection type of each actual measuring point; and querying the engine task of each component structure in the aero-engine running process in the engine running database based on the running task type corresponding to each component structure and the gas path flow direction information between the component structures.

[0076] In this embodiment, the terminal identifies the detection type of each actual measuring point based on the detection data of each actual measuring point of each component structure. The detection type includes but is not limited to outlet static pressure detection type, rotating speed detection type, outlet temperature detection type, etc. Then, the terminal queries the running task type corresponding to each component structure in the component running database based on the detection type of each actual measuring point. The component running database includes the detection type range corresponding to each running task type. The running task type is used to represent the running task of the component structure.

[0077] Then, the terminal queries the engine task of each component structure in the aero-engine operation process in the engine operation database based on the operation task type corresponding to each component structure and the gas path flow direction information between the component structures. The engine operation database includes the corresponding relationship between different engine tasks, gas path flow direction information, and operation task types. Then, the terminal identifies the engine task of each component structure in the aero-engine operation process based on the corresponding relationship. For example, the key measurement points of the compressor include the outlet static pressure and the rotating speed, and the key measurement point of the turbine is the outlet temperature. Through the flow direction of the airflow, the compressor is located upstream of the airflow, and the outlet static pressure and the rotating speed of the compressor correspond to the gas compression task. The terminal identifies the engine task of the compressor as an upstream task by querying the engine operation database. Through the flow direction of the airflow, it can be known that the turbine is located downstream of the airflow, and the outlet temperature of the turbine corresponds to the gas transmission type. The terminal identifies the engine task of the turbine as a downstream task by querying the engine operation database. The establishment of such a task relationship reflects the causal dependence in the physical system.

[0078] Based on the above scheme, by combining the operation task type corresponding to each component structure and the gas path flow direction information between the component structures, the engine task corresponding to each component structure is identified, and the accuracy of identification is improved.

[0079] Optionally, based on the task association relationship between the component structures and the engine task corresponding to each component structure, the task dependency information between the component structures is identified, including: identifying the information transmission mode between the component structures based on the task association relationship between the component structures; identifying the input requirement information corresponding to each component structure and the output information corresponding to each component structure based on the engine task corresponding to each component structure, and identifying the data interaction information between the component structures based on the information transmission mode between the component structures, the input requirement information corresponding to each component structure, and the output information corresponding to each component structure; and taking the data interaction information between the component structures as the task dependency information between the component structures.

[0080] In this embodiment, the terminal identifies the information transmission mode between the component structures based on the task association relationship between the component structures. The information transmission mode is the data transmission mode of the actual measurement point data between the component structures.

[0081] Then, the terminal identifies the input requirement information corresponding to each component structure and the output information corresponding to each component structure based on the engine task corresponding to each component structure, and identifies the data interaction information between the component structures based on the information transmission mode between the component structures, the input requirement information corresponding to each component structure, and the output information corresponding to each component structure. The data interaction information is the data interaction logic between two adjacent component structures obtained by establishing the data interaction strategy between the input information and the output information between the two adjacent component structures in the order of the engine tasks. Finally, the terminal takes the data interaction information between the component structures as the task dependency information between the component structures.

[0082] Based on the above scheme, the information transmission between different tasks is optimized by constructing a reasonable task dependency relationship, which significantly reduces the complexity of directly using an end-to-end neural network for learning. By modeling the dependency between tasks, the model can better capture the correlation between tasks, thereby improving the accuracy and reliability of engine performance prediction.

[0083] Optionally, based on the flight condition information and the engine tasks corresponding to the component structures, a low-fidelity network is constructed, including: obtaining baseline simulation data of the aero-engine; identifying each flight condition data in the flight condition information, and training an initial low-fidelity network based on each baseline simulation data and each flight condition data to obtain the low-fidelity network.

[0084] In this embodiment, the terminal obtains baseline simulation data of the aero-engine. Then, the terminal identifies each flight condition data in the flight condition information, and trains an initial low-fidelity network based on each baseline simulation data and each flight condition data to obtain the low-fidelity network. Specifically, the low-fidelity network NN L Based on the engine baseline simulation data, the low-fidelity network is trained to approximate the data model under the baseline condition, and the input of the low-fidelity network is the flight condition P 0L ,T 0L ,W fL , and the output is the predicted model parameter Y L .

[0085] Based on the above scheme, the model parameter setting accuracy of the data-driven model is improved by training the low-fidelity network.

[0086] Optionally, based on the low-fidelity network, the flight condition information, and the task dependency information between the component structures, a high-fidelity network is constructed, including: based on the data interaction information between the component structures, identifying sub-data interaction information between the detection data of each actual measuring point, and based on each flight condition data, obtaining each model parameter through the low-fidelity network; based on the detection data of each actual measuring point, each flight condition data, and each model parameter, training an initial high-fidelity network to obtain the high-fidelity network.

[0087] In this embodiment, the terminal identifies sub-data interaction information between the detection data of each actual measuring point based on the data interaction information between the component structures, and obtains each model parameter through the low-fidelity network based on each flight condition data. The model parameters are model parameters of the data-driven model. Then, the terminal trains an initial high-fidelity network based on the detection data of each actual measuring point, each flight condition data, and each model parameter to obtain the high-fidelity network. Specifically, the high-fidelity network is composed of two sub-networks, which are used to approximate the linear and nonlinear correlations between the flight conditions and the engine measuring points in the real data, and the inputs are the flight conditions P Hnl , the model parameters Y Hn predicted by the low-fidelity network, and the outputs are the engine measuring points Y 0H predicted by the high-fidelity network. 0H fH L H

[0088] Based on the above scheme, by combining low-fidelity and high-fidelity data sources, the proposed multi-task cascaded MFPINN method can significantly improve the prediction accuracy of the model while ensuring the calculation efficiency. The model structure not only enhances the interpretability of the model, but also reduces the dependence on high-cost operating data, thereby ensuring the efficiency and economy of the model. The multi-fidelity architecture of MFPINN can effectively utilize different quality data sources and combine the prior knowledge of the engine physical system to further optimize the overall performance of the model. The method passes the prediction results of the low-fidelity network to the high-fidelity network through the multi-task cascading manner, ensuring the information flow and collaborative optimization between different network levels.

[0089] ​​​​Optionally, an engine performance model for aero-engines is constructed based on low-fidelity networks, high-fidelity networks, and airflow direction information between various component structures. This includes: acquiring physical information of each component structure of the aero-engine and constructing a physical model of the aero-engine based on the physical information of each component structure; constructing a data-driven model of the aero-engine based on low-fidelity networks and high-fidelity networks through a multi-task cascade modeling strategy; and using the physical model and the data-driven model as the engine performance model of the aero-engine.

[0090] In this embodiment, the terminal acquires the physical information of the structure of each component of the aero-engine and constructs a physical model of the aero-engine based on this information. Then, the terminal constructs a data-driven model of the aero-engine using a multi-task cascaded modeling strategy, based on both low-fidelity and high-fidelity networks. Finally, the terminal uses the physical model and the data-driven model as the engine performance model of the aero-engine. The multi-task cascaded modeling strategy is a construction strategy that uses a multi-task cascaded model to build the data-driven model.

[0091] In practical applications, within a nonlinear dynamic physical model, a series of component parameters, such as the mass flow rate and efficiency of typical components, are adjusted to simulate the mismatch between the actual engine and the model. Typical adjustable component parameters in the physical model include the efficiency and flow capacity of each component, pressure drop in pipelines, and combustion efficiency. During implementation, the efficiency and flow rate of the baseline model are assumed to always be 1; the coefficients of various factors are changed to simulate a real engine.

[0092] The simulated flight plan is as follows Figure 3 As shown, simulation results of the physical model (PBM) and the actual operating engine (ROE) were obtained through flight condition simulation. Figure 4 The data shows some of the measurable parameters N1 and N2 from the flight plan, represented as normalized results. The root mean square error (RMSE) for each parameter is greater than 2% and less than 6%, indicating that the mismatch is acceptable for the same engine type.

[0093] The MFPINN network was trained using real data from the flight plan and PBM simulation data from multiple flight plans. Simultaneously, the trained network was tested using a second set of flight plans. The second set of flight plans is as follows: Figure 5 As shown. Figure 6 The comparison between the final predicted results and the actual results is shown. As can be seen from the figure, the proposed method can significantly reduce the error of the physical model.

[0094] Based on the above scheme, by integrating the physical model and the data-driven model, the error problem between the traditional model and the actual measurement is overcome, and higher prediction accuracy and correction efficiency are achieved.

[0095] The application also provides a cascade modeling example of aero-engine multi-task learning, as shown in Figure 7 The specific processing process includes the following steps:

[0096] Step S701, obtaining detection data of each actual measuring point of each component structure of the aero-engine, task association relationship between each component structure, air path flow direction information between each component structure, and flight condition information of the aero-engine.

[0097] Step S702, identifying the detection type of each actual measuring point based on the detection data of each actual measuring point of each component structure.

[0098] Step S703, querying the running task type corresponding to each component structure in the component running database based on the detection type of each actual measuring point.

[0099] Step S704, querying the engine task of each component structure in the aero-engine running process in the engine running database based on the running task type corresponding to each component structure and the air path flow direction information between each component structure.

[0100] Step S705, identifying the information transmission mode between each component structure based on the task association relationship between each component structure.

[0101] Step S706, identifying the input requirement information corresponding to each component structure and the output information corresponding to each component structure based on the engine task corresponding to each component structure, and identifying the data interaction information between each component structure based on the information transmission mode between each component structure, the input requirement information corresponding to each component structure, and the output information corresponding to each component structure.

[0102] Step S707, taking the data interaction information between each component structure as the task dependency information between each component structure.

[0103] Step S708, obtaining baseline simulation data of the aero-engine.

[0104] Step S709, identifying each flight condition data in the flight condition information, and training an initial low-fidelity network based on each baseline simulation data and each flight condition data to obtain a low-fidelity network.

[0105] Step S710, identifying sub-data interaction information between the detection data of each actual measuring point based on the data interaction information between each component structure, and obtaining each model parameter through the low-fidelity network based on each flight condition data.

[0106] Step S711: Based on the detection data of each actual measurement point, the flight condition data, and the model parameters, train the initial high-fidelity network to obtain the high-fidelity network.

[0107] Step S712: Obtain the physical information of the structure of each component of the aero-engine, and construct a physical model of the aero-engine based on the physical information of each component.

[0108] Step S713: Based on low-fidelity networks and high-fidelity networks, a data-driven model of the aero-engine is constructed through a multi-task cascade modeling strategy.

[0109] Step S714: Use the physical model and the data-driven model as the engine performance model of the aero-engine.

[0110] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0111] Based on the same inventive concept, this application also provides a cascade modeling apparatus for implementing the cascade modeling method for multi-task learning of aero-engines as described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method. Therefore, the specific limitations in one or more embodiments of the cascade modeling apparatus for multi-task learning of aero-engines provided below can be found in the limitations of the cascade modeling method for multi-task learning of aero-engines described above, and will not be repeated here.

[0112] In one exemplary embodiment, such as Figure 8 As shown, a cascaded modeling device for multi-task learning of aero-engines is provided, comprising: an acquisition module 810, a recognition module 820, and a construction module 830, wherein:

[0113] The acquisition module 810 is configured to acquire detection data of each actual measuring point of each component structure of an aero-engine, task association relationships between the component structures, air path flow direction information between the component structures, and flight condition information of the aero-engine, and generate an engine task corresponding to each component structure based on the detection data of each actual measuring point of each component structure and the air path flow direction information between the component structures.

[0114] The identification module 820 is configured to identify task dependency information between the component structures based on the task association relationships between the component structures and the engine task corresponding to each component structure, and construct a low-fidelity network based on the flight condition information and the engine task corresponding to each component structure.

[0115] The construction module 830 is configured to construct a high-fidelity network based on the low-fidelity network, the flight condition information, and the task dependency information between the component structures, and construct an engine performance model of the aero-engine based on the low-fidelity network, the high-fidelity network, and the air path flow direction information between the component structures.

[0116] Optionally, the acquisition module 810 is specifically configured to:

[0117] identify a detection type of each actual measuring point based on the detection data of each actual measuring point of each component structure;

[0118] query, in a component operation database, an operation task type corresponding to each component structure based on the detection type of each actual measuring point;

[0119] query, in an engine operation database, an engine task of each component structure in an aero-engine operation process based on the operation task type corresponding to each component structure and the air path flow direction information between the component structures.

[0120] Optionally, the identification module 820 is specifically configured to:

[0121] identify an information transmission mode between the component structures based on the task association relationships between the component structures;

[0122] identify input demand information corresponding to each component structure and output information corresponding to each component structure based on the engine task corresponding to each component structure, and identify data interaction information between the component structures based on the information transmission mode between the component structures, the input demand information corresponding to each component structure, and the output information corresponding to each component structure;

[0123] Data interaction information between the component structures is identified as task dependency information between the component structures.

[0124] Optionally, the identification module 820 is specifically used for:

[0125] Baseline simulation data of the aero-engine is acquired.

[0126] Each flight condition data in the flight condition information is identified, and an initial low-fidelity network is trained based on each baseline simulation data and each flight condition data, to obtain a low-fidelity network.

[0127] Optionally, the construction module 830 is specifically used for:

[0128] Based on the data interaction information between the component structures, sub-data interaction information between detection data of each actual measuring point is identified, and each model parameter is acquired through the low-fidelity network based on each flight condition data.

[0129] Based on the detection data of each actual measuring point, each flight condition data, and each model parameter, an initial high-fidelity network is trained to obtain a high-fidelity network.

[0130] Optionally, the construction module is specifically used for:

[0131] Physics information of each component structure of the aero-engine is acquired, and a physical model of the aero-engine is constructed based on the physics information of each component structure.

[0132] Based on the low-fidelity network and the high-fidelity network, a data-driven model of the aero-engine is constructed through a multi-task cascade modeling strategy.

[0133] The physical model and the data-driven model are taken as an engine performance model of the aero-engine.

[0134] Each module in the above-mentioned cascade modeling device for multi-task learning of an aero-engine can be realized by software, hardware, and a combination thereof, in whole or in part. Each module can be embedded in or independent of a processor in a computer device in a hardware form, or can be stored in a memory in a computer device in a software form, so as to be called and executed by a processor to perform operations corresponding to each module.

[0135] In one exemplary embodiment, a computer device can be provided, which can be a terminal, and an internal structure diagram thereof can be as shown in Figure 9As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with the external terminal in a wired or wireless manner. The wireless manner can be realized through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to realize an aero-engine multi-task learning cascade modeling method. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0136] Those skilled in the art can understand that, Figure 9 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0137] In one exemplary embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to realize the steps corresponding to the aero-engine multi-task learning cascade modeling method.

[0138] In one embodiment, a computer readable storage medium is provided, having a computer program stored thereon, and the computer program is executed by the processor to realize the steps corresponding to the aero-engine multi-task learning cascade modeling method.

[0139] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by the processor to realize the steps corresponding to the aero-engine multi-task learning cascade modeling method.

[0140] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0141] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program. 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 above-mentioned embodiments of each method. Any reference to memory, database or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetic variable memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided by the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided by the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0142] The technical features of the above embodiments can be combined in any way. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0143] The above-described embodiments are merely illustrative of several embodiments of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

Claims

1. A cascaded modeling method for multi-task learning of aero-engines, characterized in that, The method includes: The system acquires the detection data of each actual measurement point of each component structure of the aero-engine, the task association relationship between each component structure, the airflow direction information between each component structure, and the flight condition information of the aero-engine. Based on the detection data of each actual measurement point of each component structure and the airflow direction information between each component structure, the system generates the engine task corresponding to each component structure. Based on the task relationships between the various component structures and the engine tasks corresponding to each component structure, the task dependency information between the various component structures is identified. Based on the flight condition information and the engine tasks corresponding to each component structure, a low-fidelity network is constructed. The low-fidelity network is trained based on engine baseline simulation data to approximate the data model under baseline conditions. The input data of the low-fidelity network is the flight condition information, and the output data of the low-fidelity network is the predicted model parameters of the data model. Based on the low-fidelity network, the flight condition information, and the task dependency information between the various component structures, a high-fidelity network is constructed. The high-fidelity network consists of two sub-networks, which are used to approximate the linear and nonlinear correlations between flight conditions and actual measurement points in real data, respectively. The input data of the high-fidelity network are the flight condition information and the model parameters of the data model, and the output data of the high-fidelity network is the predicted data of the predicted actual measurement points. Obtain the physical information of the structure of each component of the aero-engine, and construct a physical model of the aero-engine based on the physical information of each component structure; Based on the low-fidelity network and the high-fidelity network, a data-driven model of the aero-engine is constructed through a multi-task cascade modeling strategy. The physical model and the data-driven model are used as the engine performance model of the aero-engine.

2. The method according to claim 1, characterized in that, The engine task corresponding to each component structure is generated based on the detection data of each actual measurement point of each component structure and the airflow direction information between each component structure, including: Based on the detection data of each actual measurement point of each component structure, identify the detection type of each actual measurement point; Based on the detection type of each actual measurement point, query the running task type corresponding to each component structure in the component operation database; Based on the operational task type corresponding to each component structure and the airflow direction information between each component structure, the engine task of each component structure during the operation of the aero-engine is queried in the engine operation database.

3. The method according to claim 2, characterized in that, The process of identifying task dependency information between the component structures based on the task association relationships between each component structure and the engine tasks corresponding to each component structure includes: Based on the task association relationships between the various component structures, the information transmission methods between the various component structures are identified; Based on the engine task corresponding to each component structure, the input requirement information and output information corresponding to each component structure are identified. Based on the information transmission method between each component structure, the input requirement information and output information corresponding to each component structure, the data interaction information between each component structure is identified. The data interaction information between the various component structures is used as the task dependency information between the various component structures.

4. The method according to claim 1, characterized in that, The construction of a low-fidelity network based on the flight condition information and the engine tasks corresponding to each component structure includes: Obtain the baseline simulation data of the aero-engine; Identify each flight condition data in the flight condition information, and train an initial low-fidelity network based on each baseline simulation data and each flight condition data to obtain the low-fidelity network.

5. The method according to claim 4, characterized in that, The construction of a high-fidelity network based on the low-fidelity network, the flight condition information, and the task dependency information between the various component structures includes: Based on the data interaction information between the structure of each component, the sub-data interaction information between the detection data of each actual measurement point is identified, and based on the flight condition data, the model parameters are obtained through the low-fidelity network. Based on the detection data of each actual measurement point, the flight condition data of each, and the model parameters of each, an initial high-fidelity network is trained to obtain a high-fidelity network.

6. A cascaded modeling device for multi-task learning of aero-engines, characterized in that, The device includes: The acquisition module is used to acquire the detection data of each actual measurement point of each component structure of the aero-engine, the task association relationship between each component structure, the airflow direction information between each component structure, and the flight condition information of the aero-engine, and generate the engine task corresponding to each component structure based on the detection data of each actual measurement point of each component structure and the airflow direction information between each component structure. The identification module is used to identify the task dependency information between the component structures based on the task association relationship between each component structure and the engine task corresponding to each component structure, and to construct a low-fidelity network based on the flight condition information and the engine task corresponding to each component structure; the low-fidelity network is trained based on engine baseline simulation data to approximate the data model under baseline conditions, the input data of the low-fidelity network is the flight condition information, and the output data of the low-fidelity network is the predicted model parameters of the data model; A construction module is used to construct a high-fidelity network based on the low-fidelity network, the flight condition information, and the task dependency information between the various component structures. The high-fidelity network consists of two sub-networks, which are used to approximate the linear and nonlinear correlations between flight conditions and engine measurement points in real data, respectively. The input data of the high-fidelity network are the flight condition information and the model parameters of the data model, and the output data of the high-fidelity network is the predicted data of the actual measurement points. The module acquires the physical information of each component structure of the aero-engine and constructs a physical model of the aero-engine based on the physical information of each component structure. Based on the low-fidelity network and the high-fidelity network, a data-driven model of the aero-engine is constructed through a multi-task cascaded modeling strategy. The physical model and the data-driven model are used as the engine performance model of the aero-engine.

7. The apparatus according to claim 6, characterized in that, The acquisition module is specifically used for: Based on the detection data of each actual measurement point of each component structure, identify the detection type of each actual measurement point; Based on the detection type of each actual measurement point, query the running task type corresponding to each component structure in the component operation database; Based on the operational task type corresponding to each component structure and the airflow direction information between each component structure, the engine task of each component structure during the operation of the aero-engine is queried in the engine operation database.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images