Aero-engine multi-module adaptive coupling simulation system and method
By introducing an adaptive coupled simulation system with an AI interface model and a balanced residual controller, the problems of low iteration efficiency, error amplification, and poor adaptability to operating conditions in aero-engine simulation are solved. This achieves efficient and accurate multi-module simulation, adapting to simulation requirements under complex operating conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TAIHANG NATIONAL LABORATORY
- Filing Date
- 2026-05-06
- Publication Date
- 2026-07-14
AI Technical Summary
Existing aero-engine performance simulation methods suffer from problems such as low iteration efficiency, error propagation amplification, insufficient model matching, and poor adaptability to operating conditions. In particular, it is difficult to achieve efficient and accurate multi-module joint simulation under complex operating conditions.
A multi-module adaptive coupling simulation system is adopted. By introducing an AI interface model and a balanced residual controller, adaptive switching and convergence control between modules are realized. Combined with the conservation constraints of mass, energy and momentum, the simulation strategy and parameters are dynamically adjusted to ensure simulation accuracy and stability.
It significantly improves simulation efficiency and convergence stability, achieves physical consistency and high-precision coupling between modules, has cross-fidelity adaptive capability, adapts to simulation needs under complex working conditions, and supports modular expansion and engineering integration applications.
Smart Images
Figure CN122151554B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of aero-engine technology, and in particular to an aero-engine multi-module adaptive coupling simulation system and method. Background Technology
[0002] Aero-engine performance simulation is an indispensable and crucial step in engine design, analysis, and verification. By performing numerical simulations of the entire engine, its thermodynamic performance, component matching relationships, and energy conversion efficiency under different operating conditions can be predicted during the design phase, providing a basis for structural optimization and scheme decision-making. Traditional aero-engine performance simulation is typically based on a modular serial solution framework, which involves sequentially simulating components such as the fan, compressor, combustion chamber, turbine, and exhaust nozzle according to the airflow and energy transfer path, and achieving mass, energy, and momentum balance of the entire engine through multiple iterations. This method has been widely used in existing engineering software systems (such as NPSS and GasTurb).
[0003] However, with the increasing complexity of aero-engine structures, the diversification of fuel types, and the introduction of new combustion modes, traditional serial iterative simulation methods have gradually revealed the following limitations:
[0004] (1) Low iteration efficiency and cascaded amplification of errors: In the traditional sequential solution framework, the simulation of each component needs to be carried out in sequence according to the process. The calculation error of the upstream module will be gradually transmitted and amplified along the process, resulting in an increase in the deviation of the downstream module's results, which affects the convergence speed and stability of the overall balance solution.
[0005] (2) Insufficient physical model compatibility: The flow characteristics, thermodynamic states, and reaction mechanisms of various components of an aero-engine differ significantly. For example, the compressor is dominated by high Mach number turbulent flow, while the combustion chamber involves strongly nonlinear chemical reactions and multi-scale heat release processes. In multi-component joint simulation, traditional methods usually use the same type of turbulence and thermodynamic models under the same simulation platform or solution framework to maintain data compatibility between modules. However, this unified modeling approach is difficult to take into account the physical characteristics of different components, which can easily lead to the accumulation of overall prediction errors and local calculation deviations.
[0006] (3) Poor adaptability to operating conditions: When the engine operates at non-design points or under complex conditions (such as high-altitude restart or variable thrust adjustment), the coupling relationship between modules and boundary conditions will change significantly. Traditional static iterative models cannot dynamically adjust simulation strategies and parameters according to operating conditions, making it difficult to accurately reflect actual operating characteristics.
[0007] To address the aforementioned issues, current research has made some progress in areas such as multi-module joint simulation, digital twins, and surrogate models. However, existing methods generally suffer from limitations such as fixed coupling strategies, experience-dependent convergence control, and low computational efficiency across fidelity levels. There is still a lack of a unified simulation method that can combine artificial intelligence prediction capabilities with physical conservation constraints to achieve adaptive mapping of interfaces between modules and switching between multiple fidelities. Summary of the Invention
[0008] In view of this, embodiments of this application provide a multi-module adaptive coupling simulation system and method for aero-engines, which at least partially solves the problems of low iteration efficiency, error propagation amplification, insufficient model matching, and poor adaptability to operating conditions in existing aero-engine performance simulations. This method learns the physical mapping relationships between modules to achieve adaptive switching and convergence control between models of different fidelity for each component, thereby significantly improving the computational efficiency and accuracy of the whole-engine simulation, ensuring stable solution performance under complex operating conditions, and constructing an efficient and intelligent simulation framework for the design and analysis of next-generation aero-engines.
[0009] In a first aspect, embodiments of this application provide a multi-module adaptive coupled simulation system for aero-engines, including a multi-physics module simulation layer, an AI interface model layer, a balance residual controller, and a fidelity scheduling module. The multi-physics module simulation layer includes multiple aero-engine physical modules arranged sequentially. The AI interface model layer includes multiple AI interface models, with one AI interface model set between two adjacent aero-engine physical modules. Each aero-engine physical module and each AI interface model interacts with the balance residual controller. The output of the balance residual controller is connected to the input of the fidelity scheduling module, and the output of the fidelity scheduling module is connected to each aero-engine physical module. Each aero-engine physical module solves its parameters independently. The AI interface model is used for the physical mapping of the output parameters of the upstream aero-engine physical module and the input parameters of the downstream aero-engine physical module. The balance residual controller is used to monitor the conservation of mass, energy, and momentum of the entire engine, output conservation deviation information, and perform feedback correction on the AI interface model based on the conservation deviation information. The fidelity scheduling module is used to dynamically adjust the solution level of each aero-engine physical module according to the conservation deviation information and the state information of each aero-engine physical module.
[0010] According to a specific implementation of an embodiment of this application, the conservation deviation information includes the overall mass conservation residual, energy conservation residual, momentum conservation residual, and global conservation residual metric.
[0011] According to a specific implementation of this application, the expression for the mass conservation residual is:
[0012] ,
[0013] In the formula, R m For the residual of quality conservation, The total mass flow rate at the inlet of the entire machine. For fuel input mass flow rate, The total mass flow rate at the outlet of the entire machine;
[0014] The expression for the energy conservation residual is:
[0015] ,
[0016] In the formula, R e For the energy conservation residual, h in For the inlet specific enthalpy, h out For export enthalpy, For the lower heating value or equivalent chemical energy input per unit fuel, For shaft work commutation term, This refers to the system's heat loss term;
[0017] The expression for the momentum conservation residual is:
[0018] ,
[0019] In the formula, R p For momentum conservation residuals, in For inlet flow rate, out For export flow rate, in For inlet static pressure, out For the outlet static pressure, A in Let A be the cross-sectional area of the entrance. out For the export cross-sectional area, For external forces or equivalent additional forces;
[0020] The expression for the global conserved residual metric is:
[0021] ,
[0022] ,
[0023] Among them, R global For global conservation of residuals, , and β1, β2, and β3 are the normalized mass conservation residual, energy conservation residual, and momentum conservation residual, respectively, and the first weighting coefficient, the second weighting coefficient, and the third weighting coefficient are the third weighting coefficient, respectively.
[0024] According to a specific implementation of this application, the state information includes the calculated residual, the convergence rate, and the internal energy balance error. The calculated residual is used to characterize the degree of unsatisfiability of the discrete control equations of the current aero-engine physics module. The convergence rate is used to characterize the decay trend of the aero-engine physics module residual between adjacent iteration steps to distinguish between stable convergence, stagnant convergence, or oscillating convergence states. The internal energy balance error is used to measure the degree of closure between the input, output, and energy storage changes of the internal energy of the aero-engine physics module.
[0025] According to a specific implementation of an embodiment of this application, the step of dynamically adjusting the solution level of each aero-engine physical module based on conservation deviation information and the state information of each aero-engine physical module includes:
[0026] The various state information are normalized and merged to obtain local state variables;
[0027] A comprehensive index is obtained based on the global conservation residual metric and local state variables in the conservation deviation information.
[0028] When the comprehensive index exceeds the preset threshold, the solution level of the current aero-engine physics module is switched from low fidelity to high fidelity. Low fidelity includes zero-dimensional, one-dimensional and two-dimensional, while high fidelity includes three-dimensional.
[0029] When the comprehensive index is lower than the preset threshold, the current aero-engine physical module maintains or downgrades the solution level.
[0030] According to a specific implementation of an embodiment of this application, the fidelity scheduling module is further used to control the aero-engine physical module to perform boundary reconstruction or interpolation processing during the solution level switching process based on comprehensive indicators, so as to ensure the continuity of energy and flow field before and after the switching.
[0031] According to a specific implementation of this application, the input parameters include the pressure, temperature, flow rate, component concentration, and turbulence characteristics at the outlet of the upstream aero-engine physical module, and the output parameters include the pressure, temperature, flow rate, component concentration, and turbulence characteristics at the inlet of the downstream aero-engine physical module.
[0032] According to one specific implementation of the embodiments of this application, the physical module of the aero-engine includes a fan module, a compressor module, a combustion chamber module, a turbine module, and a tail nozzle module.
[0033] Secondly, embodiments of this application also provide a multi-module adaptive coupling simulation method for aero-engines, implemented based on the multi-module adaptive coupling simulation system for aero-engines as described in any embodiment of the first aspect, the method comprising:
[0034] Acquire engine structural parameters and operating point data, establish input parameter sets for each aero-engine physical module and define boundary conditions;
[0035] Each aero-engine physics module independently solves the initial boundary conditions and outputs a dataset for training and initializing the AI interface model, thereby training the AI interface model.
[0036] The trained AI interface model is used to map and predict parameters from the upstream outlet to the downstream inlet of the aero-engine physics module, forming a dynamic boundary input. During the prediction process, the balance residual controller calculates the conservation deviation information of the whole machine mass, energy and momentum in real time, and feeds the conservation deviation information back to the AI interface model to correct the prediction results of the AI interface model.
[0037] The fidelity scheduling module dynamically adjusts the solution level of each aero-engine physical module based on the conservation deviation information and the state information of each aero-engine physical module.
[0038] The balance residual controller continuously monitors the conservation conditions of mass, energy and momentum. When the conservation deviation information exceeds the limit, it performs corresponding feedback and correction processing on the aero-engine physics module and AI interface model respectively.
[0039] After the system converges, key performance indicators are output.
[0040] According to a specific implementation of an embodiment of this application, the correction of the prediction results of the AI interface model includes:
[0041] When the system detects a mass conservation residual, it corrects the downstream inlet mass flow rate predicted by the AI interface model.
[0042] When the system detects an energy conservation residual, it corrects the predicted total temperature or enthalpy value.
[0043] When the system detects momentum conservation residuals, it corrects the predicted total pressure or velocity parameters.
[0044] Beneficial effects:
[0045] The multi-module adaptive coupling simulation system and method for aero-engines in this application embodiment have the following beneficial effects:
[0046] Significantly improve simulation efficiency and convergence stability: By introducing an AI-based physical mapping model at the module interface, the boundary conditions of downstream components can be directly predicted, reducing information transmission and error accumulation during multiple iterations, significantly shortening the overall system balance calculation time, and improving convergence speed and computational stability.
[0047] Achieving physical consistency and high-precision coupling among multiple modules: Introducing mass, energy, and momentum conservation constraints during AI model training ensures that parameter transfer between modules conforms to physical laws, fundamentally eliminating physical incompatibility issues between different simulation software or models and improving overall computational accuracy;
[0048] It has the ability to adaptively solve across fidelity: the method automatically adjusts the model fidelity according to the uncertainty level of each module during the simulation process, and achieves a dynamic balance between global efficiency and local accuracy. It can adapt to different computing needs and complex working conditions, and is especially suitable for performance prediction under non-design point and variable fuel conditions.
[0049] Enhance the model's adaptability and generalization ability: The artificial intelligence interface mapping model can be trained on various operating condition data, has strong nonlinear mapping ability and extrapolation performance, and can maintain stable prediction accuracy when engine operating conditions change, avoiding the distortion caused by fixed parameters in traditional methods.
[0050] Supports modular expansion and engineering integration: The framework of this method is clear and the module interface is standardized, which can be seamlessly connected with existing aero-engine simulation software or digital design platforms. It is convenient to promote its application in design verification, performance evaluation and combustion optimization, etc., and has good engineering practicality and scalability.
[0051] In summary, by introducing an artificial intelligence-driven adaptive coupling mechanism, this invention achieves high efficiency, intelligence, and adaptability in multi-module performance simulation of aero-engines while ensuring physical consistency and computational accuracy, providing a new technical path for the design and combustion characteristic research of next-generation aero-engines. Attached Figure Description
[0052] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0053] Figure 1 This is a diagram of an aero-engine multi-module adaptive coupling simulation system architecture according to an embodiment of the present invention.
[0054] Figure 2 This is a flowchart of an aero-engine multi-module adaptive coupling simulation method according to an embodiment of the present invention. Detailed Implementation
[0055] The embodiments of this application will now be described in detail with reference to the accompanying drawings.
[0056] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0057] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.
[0058] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The illustrations only show the components related to this application and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.
[0059] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.
[0060] Firstly, embodiments of this application provide a multi-module adaptive coupling simulation system for aero-engines. This system combines artificial intelligence models with physical constraint solving mechanisms to achieve intelligent coupling and efficient solving among multiple components of an aero-engine. (Refer to...) Figure 1This system specifically includes a multi-physics module simulation layer, an AI interface model layer, a balance residual controller, and a fidelity scheduling module. The multi-physics module simulation layer includes multiple sequentially arranged aero-engine physics modules. The AI interface model layer includes multiple AI interface models, with one AI interface model positioned between two adjacent aero-engine physics modules. Each aero-engine physics module and each AI interface model interacts with the balance residual controller. The output of the balance residual controller is connected to the input of the fidelity scheduling module, and the output of the fidelity scheduling module is connected to each aero-engine physics module. Each aero-engine physics module solves its parameters independently. The AI interface model is used for the physical mapping of the output parameters of the upstream aero-engine physics module to the input parameters of the downstream aero-engine physics module. The balance residual controller monitors the conservation of mass, energy, and momentum of the entire engine, outputs conservation deviation information, and performs feedback correction on the AI interface model based on the conservation deviation information. The fidelity scheduling module dynamically adjusts the solution level of each aero-engine physics module according to the conservation deviation information and the state information of each aero-engine physics module.
[0061] In this embodiment, an intelligent simulation framework integrating artificial intelligence prediction, physical constraint solving and adaptive control was constructed, realizing efficient collaborative simulation of complex multi-module systems of aero-engines, and providing reliable support for performance analysis and design optimization of new combustion modes and multi-fuel aero-engines.
[0062] In practical implementation, the multi-physics module simulation layer includes components such as fans, compressors, combustion chambers, turbines, and exhaust nozzles. Each module can be solved independently using zero-dimensional, one-dimensional, two-dimensional, or three-dimensional methods. The AI interface model is used to achieve physical mapping of input-output parameters (such as pressure, temperature, and mass flow rate) between adjacent modules to ensure physical continuity between modules. Figure 1 As shown, AI interface model 1 is set between the fan module and the compressor module; AI interface model 2 is set between the compressor module and the combustion chamber module; AI interface model 3 is set between the combustion chamber module and the turbine module; and AI interface model 4 is set between the turbine module and the exhaust nozzle module. The balance residual controller is used to monitor the conservation of mass, energy, and momentum of the entire system and to implement feedback correction for each AI interface model. The fidelity scheduling module is used to dynamically adjust the solution level of a single simulation module, thereby achieving an adaptive balance between computational accuracy and efficiency. The system achieves horizontal interface parameter prediction through AI interface models, vertical hierarchical control through the fidelity scheduling module, and forms a closed-loop solution structure under the global constraints of the balance residual controller.
[0063] In one embodiment, the AI interface model, based on a machine learning or operator learning framework, learns the nonlinear mapping relationship between the upstream module outlet parameters and the downstream module inlet boundary conditions. The model's inputs include upstream outlet pressure, temperature, flow rate, component concentration, and turbulence characteristics, while the output is the physical parameters corresponding to the downstream inlet. During the training phase, mass, energy, and momentum conservation constraints are introduced to ensure the prediction results meet physical consistency. During runtime, the AI interface model receives upstream module outlet data and provides downstream inlet prediction results, while dynamically correcting based on conservation deviation information provided by the balance residual controller to reduce the impact of imbalances at the interface on the overall convergence.
[0064] Furthermore, to enhance adaptability to complex operating conditions or fuel switching scenarios, the AI interface model can output its own predicted confidence or uncertainty indicators for system monitoring and diagnosis, but it is not used as a direct criterion for the fidelity scheduling module.
[0065] In one embodiment, a balanced residual controller is used to periodically calculate the conservation residuals of the entire system's mass, energy, and momentum during simulation, preventing the accumulation of multi-module coupling errors that could lead to numerical divergence. The basic idea is to uniformly map the flow rate, enthalpy flow, and momentum flux at the boundaries of each module onto system-level conservation relationships, quantifying the overall closure error. The conservation deviation information includes the entire system's mass conservation residual, energy conservation residual, momentum conservation residual, and a global conservation residual metric.
[0066] Furthermore, the expression for the mass conservation residual is:
[0067] ,
[0068] In the formula, R m For the residual of quality conservation, The total mass flow rate at the inlet of the entire machine. For fuel input mass flow rate, The total mass flow rate at the outlet of the entire machine;
[0069] The expression for the energy conservation residual is:
[0070] ,
[0071] In the formula, R e For the energy conservation residual, h in For the inlet specific enthalpy, h out For export enthalpy, For the lower heating value or equivalent chemical energy input per unit fuel, For shaft work commutation term, This refers to the system's heat loss term;
[0072] The expression for the momentum conservation residual is:
[0073] ,
[0074] In the formula, R p For momentum conservation residuals, in For inlet flow rate, out For export flow rate, in For inlet static pressure, out For the outlet static pressure, A in Let A be the cross-sectional area of the entrance. out For the export cross-sectional area, For external forces or equivalent additional forces;
[0075] To facilitate unified judgment, the three types of residuals are dimensionless and then constructed into a globally conserved residual metric, the expression of which is:
[0076] ,
[0077] ,
[0078] Among them, R global For global conservation of residuals, , and β1, β2, and β3 are the normalized mass conservation residual, energy conservation residual, and momentum conservation residual, respectively, and the first weighting coefficient, the second weighting coefficient, and the third weighting coefficient are the third weighting coefficient, respectively.
[0079] In practice, when any conservation deviation exceeds a set threshold, the balance residual controller feeds back control information to the relevant simulation module (aero-engine physics module) to adjust the solution boundary or trigger a local re-solution. Simultaneously, it transmits mass, energy, and momentum conservation residual information and its global conservation residual metric to the corresponding AI interface model to correct interface parameter predictions and reduce inconsistencies at the interface. At the same time, the balance residual controller derives the global conservation residual metric for the fidelity scheduling module to reference, forming the global part of its comprehensive criterion. The interaction between the balance residual controller and the simulation module, and the AI interface model, is manifested as a control-feedback and data correction loop, respectively; the interaction with the fidelity scheduling module is only a unidirectional relationship providing residual metrics and does not constitute direct control.
[0080] In one embodiment, the fidelity scheduling module performs real-time evaluation of the solution accuracy and stability of individual simulation modules and automatically adjusts the solution level based on the evaluation results. This module includes three functions: state monitoring, model switching, and consistency correction. First, the state monitoring function collects the module's state information. This state information includes computational residuals, convergence rate, and internal energy balance error. The computational residuals characterize the degree of unsatisfaction of the discrete control equations of the current aero-engine physics module; the convergence rate characterizes the decay trend of the aero-engine physics module's residuals between adjacent iteration steps to distinguish between stable convergence, stagnant convergence, or oscillating convergence states; the internal energy balance error measures the degree of closure between the input, output, and energy storage changes within the aero-engine physics module to reflect the module's compliance with physical conservation principles. For unified scheduling, these three quantitative indicators can be normalized to form the local state variable σ of the i-th module. module,i , can be represented as:
[0081] ,
[0082] in, Calculate the normalized residual for the i-th module. Let be the normalized convergence rate of the i-th module. Let be the normalized internal energy balance error of the i-th module, and α1, α2 and α3 are the first coefficient, the second coefficient and the third coefficient, respectively.
[0083] Second, the model switching function is based on comprehensive indicators. Make a level selection, among which ; The globally conserved residual metric is calculated to balance the residual controller. and These are the fourth and fifth weighting coefficients, used to balance the local accuracy requirements of the module with the overall system's consistency requirements. Under typical conditions, they satisfy the following:
[0084] ,
[0085] Both can be set to fixed values during the initialization phase and can be adjusted during the simulation phase; when the local residual (the local state variable of the i-th module) dominates, increase To strengthen the focus on the accuracy of individual modules; when the global conserved residual metric dominates, increase Prioritizing the overall physical consistency of the machine. When Exceeding the preset threshold When this happens, the module is triggered to switch from low-fidelity (zero-dimensional, one-dimensional, two-dimensional) to high-fidelity (three-dimensional) solution; when When the solution level is below a preset threshold, maintain or downgrade the solution level to improve efficiency.
[0086] Third, the consistency correction function performs boundary reconstruction or interpolation processing by the simulation module during fidelity switching to ensure the continuity of energy and flow field before and after the switching, thereby maintaining numerical stability. It should be noted that the fidelity scheduling module does not directly participate in the evaluation of data transfer accuracy between simulation modules, nor does it directly interact with the AI interface model in terms of data or control. The criteria used by the fidelity scheduling module consist of two parts: one part is the local state variables directly obtained by the state monitoring unit from the corresponding simulation module, and the other part is the global residual information (global conserved residual metric) provided by the balanced residual controller. The control output of this module is only fed back to the corresponding simulation module for adjusting the solution hierarchy and convergence control.
[0087] In one embodiment, the step of dynamically adjusting the solution level of each aero-engine physics module based on conservation deviation information and the state information of each aero-engine physics module includes:
[0088] The various state information are normalized and merged to obtain local state variables;
[0089] A comprehensive index is obtained based on the global conservation residual metric and local state variables in the conservation deviation information.
[0090] When the comprehensive index exceeds the preset threshold, the solution level of the current aero-engine physics module is switched from low fidelity to high fidelity. Low fidelity includes zero-dimensional, one-dimensional and two-dimensional, while high fidelity includes three-dimensional.
[0091] When the comprehensive index is lower than the preset threshold, the current aero-engine physical module maintains or downgrades the solution level.
[0092] In one embodiment, the fidelity scheduling module is also used to control the aero-engine physical module to perform boundary reconstruction or interpolation processing during the solution level switching process based on comprehensive indicators, so as to ensure the continuity of energy and flow field before and after the switching.
[0093] In one embodiment, the input parameters include the pressure, temperature, flow rate, component concentration, and turbulence characteristics at the outlet of the upstream aero-engine physics module, and the output parameters include the pressure, temperature, flow rate, component concentration, and turbulence characteristics at the inlet of the downstream aero-engine physics module.
[0094] In one embodiment, the aero-engine physical module includes a fan module, a compressor module, a combustion chamber module, a turbine module, and a tailpipe module.
[0095] Secondly, embodiments of this application also provide a multi-module adaptive coupling simulation method for aero-engines, implemented based on the multi-module adaptive coupling simulation system for aero-engines as described in any embodiment of the first aspect, with reference to... Figure 2 The method includes:
[0096] Step 1, Module Parameter Acquisition and Operating Condition Definition: Acquire engine structural parameters and operating point data, establish input parameter sets for each aero-engine physical module, and define boundary conditions;
[0097] Step 2, Independent Module Simulation and Interface Data Acquisition: Each aero-engine physics module independently solves the preliminary boundary conditions and outputs a dataset for AI interface model training and initialization, and then trains the AI interface model;
[0098] Step 3, AI Interface Modeling and Prediction: The trained AI interface model is used to map and predict the parameters from the upstream outlet to the downstream inlet of the aero-engine physics module, forming a dynamic boundary input. During the prediction process, the balance residual controller calculates the conservation deviation information of the whole machine mass, energy and momentum in real time, and feeds the conservation deviation information back to the AI interface model to correct the prediction results of the AI interface model.
[0099] Step 4, Adaptive Fidelity Solving: The fidelity scheduling module dynamically adjusts the solution level of each aero-engine physical module based on the conservation deviation information and the state information of each aero-engine physical module, so as to achieve an adaptive balance between efficiency and accuracy.
[0100] Step 5, Balance Residual Monitoring and Feedback Control: The balance residual controller continuously monitors the conservation conditions of mass, energy and momentum. When the conservation deviation information exceeds the limit, corresponding feedback and correction processing is performed on the aero-engine physics module and AI interface model respectively.
[0101] Step 6, Results Output and Performance Evaluation: After the system converges, key performance indicators are output, such as thrust, fuel consumption rate, combustion efficiency, and outlet temperature distribution.
[0102] Furthermore, the correction of the prediction results of the AI interface model includes:
[0103] When the system detects a mass conservation residual, it corrects the downstream inlet mass flow rate predicted by the AI interface model.
[0104] When the system detects an energy conservation residual, it corrects the predicted total temperature or enthalpy value.
[0105] When the system detects momentum conservation residuals, it corrects the predicted total pressure or velocity parameters.
[0106] The corrected input parameters can be expressed as:
[0107] in, The input parameter vector predicted by the AI interface model. To correspond to the conserved residual, This is to correct the coefficient matrix. Through this correction process, while maintaining the prediction efficiency of the AI interface, it is ensured that the parameters transferred between modules meet the overall system conservation constraints, thereby improving the stability and physical consistency of the coupled simulation. By introducing an interface parameter correction mechanism based on conservation residual feedback, the synergistic coupling of the artificial intelligence prediction model and the physical conservation constraints is achieved.
[0108] The embodiments provided by the present invention have the following advantages:
[0109] Significantly improve simulation efficiency and convergence stability: By introducing an AI-based physical mapping model at the module interface, the boundary conditions of downstream components can be directly predicted, reducing information transmission and error accumulation during multiple iterations, significantly shortening the overall system balance calculation time, and improving convergence speed and computational stability.
[0110] Achieving physical consistency and high-precision coupling among multiple modules: Introducing mass, energy, and momentum conservation constraints during AI model training ensures that parameter transfer between modules conforms to physical laws, fundamentally eliminating physical incompatibility issues between different simulation software or models and improving overall computational accuracy;
[0111] It has the ability to adaptively solve across fidelity: the method automatically adjusts the model fidelity according to the uncertainty level of each module during the simulation process, and achieves a dynamic balance between global efficiency and local accuracy. It can adapt to different computing needs and complex working conditions, and is especially suitable for performance prediction under non-design point and variable fuel conditions.
[0112] Enhance the model's adaptability and generalization ability: The artificial intelligence interface mapping model can be trained on various operating condition data, has strong nonlinear mapping ability and extrapolation performance, and can maintain stable prediction accuracy when engine operating conditions change, avoiding the distortion caused by fixed parameters in traditional methods.
[0113] Supports modular expansion and engineering integration: The framework of this method is clear and the module interface is standardized, which can be seamlessly connected with existing aero-engine simulation software or digital design platforms. It is convenient to promote its application in design verification, performance evaluation and combustion optimization, etc., and has good engineering practicality and scalability.
[0114] In summary, by introducing an artificial intelligence-driven adaptive coupling mechanism, this invention achieves high efficiency, intelligence, and adaptability in multi-module performance simulation of aero-engines while ensuring physical consistency and computational accuracy, providing a new technical path for the design and combustion characteristic research of next-generation aero-engines.
[0115] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A multi-module adaptive coupling simulation system for aero-engines, characterized in that, The system comprises a multi-physics module simulation layer, an AI interface model layer, a balance residual controller, and a fidelity scheduling module. The multi-physics module simulation layer includes multiple sequentially arranged aero-engine physics modules. The AI interface model layer includes multiple AI interface models, with one AI interface model positioned between adjacent aero-engine physics modules. Each aero-engine physics module and each AI interface model interacts with the balance residual controller. The output of the balance residual controller is connected to the input of the fidelity scheduling module, and the output of the fidelity scheduling module is connected to each aero-engine physics module. Each aero-engine physics module solves its parameters independently. The AI interface model is used for the physical mapping of the output parameters of the upstream aero-engine physics module to the input parameters of the downstream aero-engine physics module. The balance residual controller monitors the conservation of mass, energy, and momentum of the entire engine, outputting conservation deviation information and performing feedback correction on the AI interface models based on this deviation information. The fidelity scheduling module dynamically adjusts the solution level of each aero-engine physics module based on the conservation deviation information and the state information of each aero-engine physics module. The conservation deviation information includes the overall mass conservation residual, energy conservation residual, momentum conservation residual, and global conservation residual measurement; The expression for the mass conservation residual is: ， In the formula, R m For the residual of quality conservation, The total mass flow rate at the inlet of the entire machine. For fuel input mass flow rate, The total mass flow rate at the outlet of the entire machine; The expression for the energy conservation residual is: ， In the formula, R e For the energy conservation residual, h in For the inlet specific enthalpy, h out For export enthalpy, For the lower heating value or equivalent chemical energy input per unit fuel, For shaft work commutation term, This refers to the system's heat loss term; The expression for the momentum conservation residual is: ， In the formula, R p For momentum conservation residuals, in For inlet flow rate, out For export flow rate, in For inlet static pressure, out For the outlet static pressure, A in Let A be the cross-sectional area of the entrance. out For the export cross-sectional area, For external forces or equivalent additional forces; The expression for the global conserved residual metric is: ， ， Among them, R global For global conservation of residuals, , and β1, β2, and β3 are the normalized mass conservation residual, energy conservation residual, and momentum conservation residual, respectively, and the first weighting coefficient, the second weighting coefficient, and the third weighting coefficient are the third weighting coefficient, respectively.
2. The aero-engine multi-module adaptive coupling simulation system according to claim 1, characterized in that, The state information includes the calculated residual, convergence rate, and internal energy balance error. The calculated residual is used to characterize the degree of unsatisfaction of the discrete control equations of the current aero-engine physics module. The convergence rate is used to characterize the decay trend of the aero-engine physics module residual between adjacent iteration steps to distinguish between stable convergence, stagnant convergence, or oscillating convergence states. The internal energy balance error is used to measure the degree of closure between the input, output, and energy storage changes of the internal energy of the aero-engine physics module.
3. The aero-engine multi-module adaptive coupling simulation system according to claim 2, characterized in that, The dynamic adjustment of the solution hierarchy for each aero-engine physics module based on conservation deviation information and the state information of each aero-engine physics module includes: The various state information are normalized and merged to obtain local state variables; A comprehensive index is obtained based on the global conservation residual metric and local state variables in the conservation deviation information. When the comprehensive index exceeds the preset threshold, the solution level of the current aero-engine physics module is switched from low fidelity to high fidelity. Low fidelity includes zero-dimensional, one-dimensional and two-dimensional, while high fidelity includes three-dimensional. When the comprehensive index is lower than the preset threshold, the current aero-engine physical module maintains or downgrades the solution level.
4. The aero-engine multi-module adaptive coupling simulation system according to claim 3, characterized in that, The fidelity scheduling module is also used to control the aero-engine physical module to perform boundary reconstruction or interpolation processing during the solution level switching process based on comprehensive indicators, so as to ensure the continuity of energy and flow field before and after the switching.
5. The aero-engine multi-module adaptive coupling simulation system according to claim 1, characterized in that, The input parameters include the pressure, temperature, flow rate, component concentration, and turbulence characteristics at the outlet of the upstream aero-engine physics module, while the output parameters include the pressure, temperature, flow rate, component concentration, and turbulence characteristics at the inlet of the downstream aero-engine physics module.
6. The aero-engine multi-module adaptive coupling simulation system according to claim 1, characterized in that, The physical modules of an aero-engine include the fan module, compressor module, combustion chamber module, turbine module, and exhaust nozzle module.
7. A multi-module adaptive coupling simulation method for aero-engines, implemented based on the multi-module adaptive coupling simulation system for aero-engines as described in any one of claims 1-6, characterized in that, include: Acquire engine structural parameters and operating point data, establish input parameter sets for each aero-engine physical module and define boundary conditions; Each aero-engine physics module independently solves the initial boundary conditions and outputs a dataset for training and initializing the AI interface model, thereby training the AI interface model. The trained AI interface model is used to map and predict parameters from the upstream outlet to the downstream inlet of the aero-engine physics module, forming a dynamic boundary input. During the prediction process, the balance residual controller calculates the conservation deviation information of the whole machine mass, energy and momentum in real time, and feeds the conservation deviation information back to the AI interface model to correct the prediction results of the AI interface model. The fidelity scheduling module dynamically adjusts the solution level of each aero-engine physical module based on the conservation deviation information and the state information of each aero-engine physical module. The balance residual controller continuously monitors the conservation conditions of mass, energy and momentum. When the conservation deviation information exceeds the limit, it performs corresponding feedback and correction processing on the aero-engine physics module and AI interface model respectively. After the system converges, key performance indicators are output.
8. The aero-engine multi-module adaptive coupling simulation method according to claim 7, characterized in that, The correction of the prediction results of the AI interface model includes: When the system detects a mass conservation residual, it corrects the downstream inlet mass flow rate predicted by the AI interface model. When the system detects an energy conservation residual, it corrects the predicted total temperature or enthalpy value. When the system detects momentum conservation residuals, it corrects the predicted total pressure or velocity parameters.