End-to-end design methods, systems, equipment, and storage media for aero-engines
By adopting the end-to-end integrated design method for aero-engines, the problems of data silos and collaboration in the design and manufacturing process of aero-engines have been solved. This has enabled bidirectional data connectivity across the entire design and manufacturing process, improved the efficiency of fault tracing and development, and formed a multi-objective optimization experience base.
Patent Information
- Application Number
- CN202611537127.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-30
- Publication Date
- 2026-11-10
AI Technical Summary
Existing technologies for aero-engine design and manufacturing suffer from problems such as data silos across the entire chain, design and manufacturing collaboration limited to one-way data transfer, lengthy multidisciplinary iteration cycles, and the inability to reuse design experience. These issues result in long troubleshooting cycles, high rework costs, and wasted computing resources.
By adopting the full-link integrated design method for aero-engines, and through the integrated linkage of component underlying data models, we can achieve layered decomposition of requirements and bidirectional model linkage, parallel multidisciplinary collaborative design of components, automatic mapping and verification of design and process, and establish a four-layer unified data model to achieve bidirectional data consistency across the entire design-process-test chain.
It improved data consistency and fault tracing efficiency, shortened the development cycle, reduced trial production rework, formed a reusable multi-objective optimization experience library, and realized full-link collaborative optimization of design and manufacturing.
Smart Images

Figure CN123069554A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of digital design and systems engineering technology for aero-engines, and in particular, to a method for end-to-end integrated design of aero-engines. Furthermore, this invention also relates to a system, device, and storage medium incorporating the aforementioned end-to-end integrated design method for aero-engines. Background Technology
[0002] Aero-engines are complex systems engineering products involving deep interdisciplinary collaborations such as fluid mechanics, heat transfer, solid mechanics, combustion science, control theory, and materials science. The development of a modern aero-engine typically involves multiple stages, including requirements analysis, overall design, component aerodynamic design, detailed structural design, multidisciplinary simulation verification, process planning, component prototyping, component testing, full-engine bench testing, high-altitude simulation testing, and flight verification. It involves the collaborative work of multiple entities, including the overall design team, component design team, multidisciplinary simulation team, process team, and testing team, and the development cycle can last for several years or even more than ten years.
[0003] In the forward development of aero engines, there are four common technological bottlenecks across the industry: 1) The problem of data silos, unidirectional flow, and traceability failures across the entire chain; traditional MBSE (Model-Based Systems Engineering) platforms and PLM (Product Lifecycle Management) systems only cover partial R&D stages. Requirements, multidisciplinary simulation, 3D geometry, process, and test data belong to independent systems, lacking a unified data source for connection between them; requirements data is stored in the requirements management system, multidisciplinary simulation data is scattered in different professional tools such as CFD (Computational Fluid Dynamics), finite element, and rotor dynamics, 3D geometric models are stored in CAD systems, process data is stored in CAM / process management systems, and test data is stored in test data management systems; downstream process and test data cannot automatically drive upstream design, and there is a lack of standardized end-to-end traceability for requirements changes and model modifications; when index deviations occur, designers cannot quickly locate the source of the original constraints, resulting in long troubleshooting cycles and high rework costs; research shows that the core value of the digital thread lies in "using data as the thread to connect different stages of the product lifecycle," however, this concept has not yet been fully implemented in the field of aero-engines in the current technology; 2) Design and manufacturing are delayed in collaboration, manufacturability risks emerge late, and rework costs are high. In the traditional R&D process, process engineers only intervene in process review after the 3D drawings are frozen. Manufacturing feasibility factors such as machining tolerances, clamping schemes, and forming constraints are not considered simultaneously in the aerodynamic and structural design stages. There is a lack of automated two-way verification closed loop between the design and process ends. A large number of casting defects, welding deformations, and high-precision blade machining deviations are only exposed in the prototype trial production stage, leading to a series of problems such as repeated drawing revisions, part scrapping, and mold remaking, resulting in a surge in development costs. Although digital lead-ahead technology has been proposed for "data-driven, model-based analysis and decision-making to support system lifecycle activities," its current applications are still mainly concentrated in the design stage, and data integration between design and manufacturing has not yet been automated. 3) Multidisciplinary iterative fully manual driven optimization mechanism without hierarchical deviation; existing solutions only support global unified iteration and cannot distinguish which level of index deviation comes from the top-level overall constraints, component aerodynamic / structural design, or part machining tolerance; every time the performance fails to meet the standard, the entire process needs to be recalculated, resulting in huge computational load and lengthy iteration cycle for multidisciplinary coupled simulation; in the field of multidisciplinary design optimization (MDO), although methods such as Pareto optimal solution set have been studied, they lack deep integration with the full-link data model and cannot quickly output multi-objective optimal solutions; 4) Fragmented multi-disciplinary tools and inability to accumulate and reuse design experience; software interfaces such as aerodynamic CFD, rotor dynamics, combustion simulation, and process planning are not interoperable, and boundary conditions need to be manually transferred; the component design experience and multi-objective optimization experience of mature models lack standardized storage carriers, and new models repeatedly carry out a large number of repetitive simulation calculations, making it difficult to simultaneously balance multiple dimensions of objectives such as thrust, fuel consumption, life, weight, and manufacturing cost.
[0004] In existing technologies, the main design platforms used fall into the following three categories: 1) A multidisciplinary collaborative design platform for aero-engine overall systems and components based on MBSE; this scheme constructs a lightweight digital mainline (Digital T system, equipment, storage medium read), covering the first half of the design chain from top-level requirement decomposition to system architecture modeling, overall multidisciplinary simulation, and component 3D modeling; related research has proposed a "Requirement-Function-Logic-Physical" (RFLP) systems engineering framework and developed an efficient decomposition and synthesis model for overall systems and components; other research has constructed a forward design process for aero-engines based on the systems engineering V model, unfolding from requirement definition, requirement analysis, requirement confirmation, functional analysis, logical architecture design, and physical architecture design; however, the digital mainline of this scheme stops at the component simulation stage, lacking a two-way closed loop of process and testing; there is no hierarchical automatic optimization iteration mechanism, the system cannot automatically determine whether the simulation index deviation belongs to the overall level or the component level, and cannot achieve accurate local iterative optimization; 2) PLM integrated R&D and production collaboration platform for aviation equipment; This solution uses PLM product lifecycle management as its underlying foundation, opening up a one-way transformation link from design to process, covering the design, process, and production trial stages, and is currently the most widely used R&D and production collaboration solution in aero-engine enterprises; However, this solution completely lacks four core modules: top-level requirement decomposition, automatic optimization of multi-disciplinary overall solutions, digital twin test virtual and real verification, and hierarchical deviation iteration; data flows in a one-way manner, and downstream R&D stages cannot achieve reverse constraint on upstream design; 3) A lifecycle management approach for aerospace equipment based on digital threads; Existing solutions propose a lifecycle management approach for aerospace equipment based on digital threads (Digital T system, equipment, storage medium read). By acquiring process parameters at each stage, the master data of the digital threads and associated test sequences are determined, and the deviation between the physical equipment state and the virtual simulation state is mapped. The digital thread is considered a "game-changing disruptive technology." However, the digital thread architecture of this solution is a two-layer structure, and it has not established a four-layer integrated associated data model covering requirements, functions, performance, and physics; it does not include a two-way closed-loop verification mechanism for design-process pre-process; and it has not implemented local iterative optimization driven by layered deviations. In summary, existing technical solutions suffer from several shortcomings in the digital pipeline. Backend processes and experimental data cannot automatically feed back into the frontend design. The collaboration between design and manufacturing processes remains at the level of one-way data transmission. These solutions are all localized or partial links. Furthermore, they cannot distinguish the hierarchical attribution of indicator deviations. Global recalculation leads to a waste of computing resources and a lengthy iteration cycle. Summary of the Invention
[0005] This invention provides a complete end-to-end design method, system, equipment, and storage medium for aero-engines to address the technical problems in existing design methods, such as the existence of breakpoints in the digital mainline, the inability of back-end process and test data to automatically feed back to the front-end design, the reliance on one-way data transmission for design and manufacturing process collaboration, the existence of solutions for local links or partial links, the inability to distinguish the hierarchical attribution of index deviations, and the waste of computing resources and lengthy iteration cycles caused by global recalculation.
[0006] According to one aspect of the present invention, a method for end-to-end design of an aero-engine is provided, comprising: S1. The integrated underlying data model of components includes related design requirement models, functional models, performance models and physical models. All data in the model is synchronized bidirectionally to a unified digital mainline in real time. S2. Decomposition of requirements and bidirectional model linkage; Receive external whole machine user requirements, call the design requirement model and functional model to complete the three-level layer-by-layer decomposition of system, component and part, establish a four-dimensional traceability mapping relationship, and synchronously update to the underlying data model of step S1. S3. Parallel multidisciplinary collaborative design of components and optimal solution set solving; including distributed parallel simulation, Pareto optimal solution set solving and bidirectional write-back; S4. Component engineering drawings and physical geometry modeling correction; S5. Automatic mapping and verification of the design process: identify all machining geometric features in the physical model established in step S4 and generate process routes; automatically verify whether the feasibility of the process meets the preset threshold; if the threshold is not met, automatically generate quantitative optimization constraints and backflow them to step S4 through the digital main line.
[0007] As a further improvement to the above technical solution, the design requires a structured model that stores the qualitative and quantitative requirements of the entire machine, three-level quantitative constraints at the system, component, and part levels, rules for the transmission of requirement changes, and original logs for full-link traceability. The qualitative and quantitative requirements include top-level thrust, fuel consumption, lifespan, weight, and cost. The functional model carries the functional definitions, functional logic mappings, and functional index decomposition relationships of the entire machine's subsystems, supporting the automatic transmission of requirement changes to each module. The performance model integrates a standardized parameter library of multiple disciplines, including aerodynamics, thermodynamics, combustion, structural strength, rotor vibration, and control, storing simulation input boundaries, simulation output performance results, and performance verification thresholds at each level. It supports distributed multi-field coupled joint simulation calculations and stores Pareto multi-objective optimization solution sets. The physical model stores a complete set of geometric manufacturing information, including the three-dimensional geometry of parts, assembly constraints, machining features, tolerance chains, and process references. It supports the automatic standardized mapping of the design model to the process system and synchronously receives geometric correction constraints from process backflow.
[0008] As a further improvement to the above technical solution, step S2 includes: S21. Receive external whole machine user requirements, call the design requirement model and functional model to complete the three-level layer-by-layer decomposition of system, component and part, transform qualitative requirements into calculable quantitative constraints, and output the initial whole machine overall performance scheme. S22. Establish a four-dimensional traceability mapping relationship based on the four-layer model, so that top-level requirement changes can be automatically transmitted down to the component and part levels; S23. All the decomposed quantitative indicators and functional constraints are synchronously updated to the underlying model of step S1 through the digital mainline, and a complete record of requirement changes is retained.
[0009] As a further improvement to the above technical solution, step S3 includes: S31. Read the hierarchical quantization constraints output from step S2 from the digital mainline, and adopt a distributed parallel simulation architecture to simultaneously carry out multi-disciplinary coupled calculations such as aerodynamic and thermodynamic simulation, structural strength verification, rotor vibration analysis, whole machine weight estimation, and development cost calculation for each component. S32. Solve the Pareto optimal solution set that simultaneously satisfies all constraints of thrust, fuel consumption, lifespan, weight, and manufacturing cost based on a multi-objective optimization algorithm; S33. Write back all simulation performance and function parameters of the Pareto optimal solution set obtained in step S32 to the performance model and function model in step S1 in both directions, and update the unified digital mainline data source.
[0010] As a further improvement to the above technical solution, step S4 includes: S41. Read the optimal performance parameters output in step S3, and call the physical model template in step S1 to complete the 3D engineering modeling, dimensioning, and tolerance chain allocation of the parts; S42. Receive the process quantification constraints from step S5 in real time and automatically correct the model geometry, machining tolerances, and assembly datum. S43. The corrected complete physical model data is synchronized back to the data model in step S1 to complete the version archiving.
[0011] As a further improvement to the above technical solution, step S5 includes: S51. Identify all machining geometric features within the physical model of step S4, and generate standardized machining process routes, tooling schemes, tolerance allocation strategies, and trial production cost calculation results in batches; S52. Verify whether the process feasibility, processing cost, and tolerance processing accuracy meet the preset thresholds; S53. All iterative modifications, version changes, and verification logs are uniformly stored in the underlying data model.
[0012] As a further improvement to the above technical solution, the design method also includes step S6: Experimental data is written back and compared with the simulation results in the performance model; experimental data is collected, and the results are automatically compared with the simulation results in the performance model to generate deviation optimization constraints and push them back to the design end.
[0013] According to another aspect of the present invention, a fully integrated forward design system for aero-engines is also provided, comprising: The four-layer unified data model management module is used to create, associate, read, write, and version control the underlying data model. The underlying data model includes the associated design requirement model, functional model, performance model, and physical model. The requirement layering decomposition and traceability module is used to read the design requirement model to complete the three-level requirement decomposition, build a four-dimensional bidirectional traceability link, and automatically handle the global transmission of requirement changes. The parallel multidisciplinary evaluation and multi-objective optimization module is used to drive distributed multi-field coupled simulation and solve for Pareto optimal solution sets. The engineering modeling and geometry correction module is used to complete 3D modeling based on performance parameters and automatically correct geometry and tolerances upon receiving process reflow constraints. The design-process bidirectional closed-loop verification module is used to automatically analyze geometric machining features and generate process plans, and output manufacturing optimization constraints in reverse. The test data reverse write-back module is used to access data collected from bench tests and component tests, automatically benchmark against the simulation performance model, generate deviation optimization constraints, and push them back to the design end. The full-process version traceability and retrieval module is used to uniformly archive model changes, simulation records, process iterations, and test fault logs throughout the entire process. It is also used to support tracing back from test failures to the top-level original requirement constraints.
[0014] According to another aspect of the present invention, a dedicated electronic device for forward design of aero-engines is also provided, including a processor and a memory; the memory stores the underlying data model and the full-link forward design calculation program corresponding to any of the methods described in the computer execution program; the processor loads and executes the program to complete all the steps of any of the methods described.
[0015] According to another aspect of the present invention, a computer-readable storage medium is also provided, storing a computer-executable program, which, when loaded and run by a processor, executes any of the aforementioned aero-engine end-to-end design methods.
[0016] The present invention has the following beneficial effects: This design methodology unifies the storage of four types of data—requirements, functions, performance, and geometry—that were previously isolated and belonged to different systems, into a single data architecture. This establishes strong relationships, significantly improving data consistency. It upgrades the traditional single digital thread into a four-tiered, interconnected data model encompassing design requirements, functions, performance, and physical aspects. This unifies requirements, functions, simulation, and geometry information from the data source, addressing the shortcomings of traditional digital threads, such as fragmented data classification and weak model correlation. Through structured requirement decomposition rules and an automated transmission mechanism, the completeness and consistency of requirement decomposition are ensured, preventing requirement omissions and misunderstandings. The automated transmission mechanism allows top-level requirement changes to be propagated across the entire chain in a short time. The four-layer model is tightly linked through a four-dimensional bidirectional traceability mapping relationship, allowing for seamless integration between any layer. The system automatically synchronizes changes to the other three layers of the model with the primary data, allowing any geometric feature in the final physical model to be traced back to the initial design requirements and functional specifications. When a fault occurs during commissioning, the system can quickly pinpoint the source of the fault—which level, which stage, and which original constraint—improving fault tracing efficiency by over 80%. The distributed parallel simulation architecture enables simultaneous work across components and disciplines, shortening the overall forward development cycle. Multiple equilibrium solutions are provided through Pareto optimal solutions, which can be flexibly selected according to the actual needs of the project. Furthermore, the Pareto optimal solutions are stored in the performance model, forming a reusable multi-objective optimization experience library, and the data source is updated through bidirectional write-back. The system automatically maps process schemes based on the physical model, and automatically generates quantitative optimization constraints to flow back to the physical model when manufacturing feasibility is not met. Unlike the existing one-way output from design to process and manual offline communication and modification mode, this method identifies and eliminates process risks in advance, reducing trial production rework from the source. The four-layer model data in this method is bidirectionally read and written throughout the process. The standardized reverse writing of process simulation, bench test and iteration deviation data back to the underlying data model breaks the data silos of the existing technology that only have one-way flow in the design stage and offline import of process test. Downstream development data can automatically feed back to upstream design, realizing bidirectional connection of data from the same source in the entire chain of design-process-test.
[0017] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description
[0018] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a simplified flowchart of a preferred embodiment of the present invention. Detailed Implementation
[0019] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, the present invention can be implemented in many different ways as defined and covered below.
[0020] Figure 1 This is a simplified flowchart of a preferred embodiment of the present invention.
[0021] like Figure 1 As shown, the end-to-end design method for aero-engines in this embodiment includes: S1. The integrated underlying data model of components includes related design requirement models, functional models, performance models and physical models. All data in the model is synchronized bidirectionally to a unified digital mainline in real time. S2. Decomposition of requirements and bidirectional model linkage; Receive external whole machine user requirements, call the design requirement model and functional model to complete the three-level layer-by-layer decomposition of system, component and part, establish a four-dimensional traceability mapping relationship, and synchronously update to the underlying data model of step S1. S3. Parallel multidisciplinary collaborative design of components and optimal solution set solving; including distributed parallel simulation, Pareto optimal solution set solving and bidirectional write-back; S4. Component engineering drawings and physical geometry modeling correction; S5. Automatic mapping and verification of the design process: Identify all machining geometric features in the physical model of step S4 and generate process routes. Automatically verify whether the feasibility of the process meets the preset threshold. If the threshold is not met, automatically generate quantitative optimization constraints and backflow them to step S4 through the digital main line.
[0022] Understandably, this design methodology unifies the storage of four types of data—requirements, functions, performance, and geometry—that were previously isolated and belonged to different systems, into a single data architecture. This establishes strong relationships, significantly improving data consistency. It upgrades the traditional single digital thread into a four-tiered linked data model encompassing design requirements, functions, performance, and physical aspects. This unifies requirements, functions, simulation, and geometry information from the data source, addressing the shortcomings of traditional digital threads, such as fragmented data classification and weak model correlation. Through structured requirement decomposition rules and an automated transmission mechanism, the completeness and consistency of requirement decomposition are ensured, avoiding requirement omissions and misunderstandings. The automated transmission mechanism allows top-level requirement changes to be transmitted across the entire chain in a short time. The four-layer models are closely linked through a four-dimensional bidirectional traceability mapping relationship. Any data change at any level automatically synchronizes with the other three levels of the model, allowing any geometric feature of the final physical model to be traced back to the initial design requirements and functional requirements. When a fault occurs during commissioning, the fault can be quickly located at which level, which link, and which original constraint it originated from, improving fault tracing efficiency by more than 80%. The distributed parallel simulation architecture enables various components and disciplines to work simultaneously, shortening the overall forward development cycle. Multiple equilibrium schemes are provided through Pareto optimal solution sets, which can be flexibly selected according to the actual needs of the project. The Pareto optimal solution sets are stored in the performance model, forming a reusable multi-objective optimization experience library, and the data source is updated through bidirectional write-back. Based on the physical model, the process scheme mapping is automatically completed, and when manufacturing feasibility is not met, quantitative optimization constraints are automatically generated and fed back to the physical model. Unlike the existing one-way output from design to process and manual offline communication and modification mode, this method identifies and eliminates process risks in advance, reducing trial production rework from the source. The four-layer model data in this method is bidirectionally read and written throughout the process. The standardized reverse writing of process simulation, bench test and iteration deviation data back to the underlying data model breaks the data silos of the existing technology that only have one-way flow in the design stage and offline import of process test. Downstream development data can automatically feed back to upstream design, realizing bidirectional connection of data from the same source in the entire chain of design-process-test.
[0023] In some preferred embodiments, the design requirement model stores qualitative and quantitative requirements such as top-level thrust, fuel consumption, lifespan, weight, and cost of the entire machine in a structured manner, along with three levels of quantitative constraints at the system, component, and part levels, rules for requirement change transmission, and original logs for full-link traceability. External user requirements are uniformly structured and stored in the database as benchmark indicators for the entire process. It can be understood that the design requirement model transforms unstructured or semi-structured requirements provided by users (such as "engine thrust not less than XX kN" and "fuel consumption not higher than XX g / (kN·s)") into a structured, calculable, and traceable quantitative constraint system. The design requirement model is the source and benchmark of the entire design process. All subsequent functional definitions, performance simulations, and geometric designs are based on the quantitative constraints in the design requirement model as the starting point and judgment basis. When the top-level requirements change, the model automatically triggers the change transmission mechanism to ensure that the change information is transmitted to all relevant levels. The functional model carries the functional definitions, functional logic mappings, and functional indicator decomposition relationships of the whole machine subsystems. It binds the hierarchical relationship of "top-level requirement indicators - subsystem functions - component functions" to support the automatic transmission of requirement changes to various professional functional modules. The functional model transforms the quantitative constraints in the design requirement model into the functional requirements of each subsystem and component. For example, the top-level thrust requirement is decomposed into compressor pressure ratio functional requirements, turbine expansion ratio functional requirements, combustion room temperature rise functional requirements, etc. The model establishes a mapping bridge between "requirement and function" to ensure that each design requirement has a corresponding functional implementation plan, avoiding requirement suspension and functional omission. The performance model integrates standardized parameter libraries from multiple disciplines, including aerodynamics and thermodynamics, combustion, structural strength, rotor vibration, and control. It stores simulation input boundaries, simulation output performance results, and performance verification thresholds at each level, supporting distributed multi-field coupled joint simulation calculations and storing Pareto multi-objective optimization solution sets. As a data center for multi-disciplinary simulations, the performance model uniformly manages the input boundary conditions and output results of various professional simulation tools. In traditional solutions, aerodynamic CFD, thermal analysis, and structural strength simulation tools run independently, with inconsistent data formats and incompatible interfaces. The performance model, through a standardized parameter library and bidirectional API interfaces, achieves unified storage and automatic exchange of multi-disciplinary simulation data, providing a data foundation for distributed parallel simulation and multi-objective optimization. The physical model stores a complete set of geometric manufacturing information for parts, including 3D geometry, assembly constraints, machining features, tolerance chains, and process datums. It supports the automatic standardized mapping of the design model to the process system and synchronously receives geometric correction constraints from process backflow. As the geometric carrier of the digital twin, the physical model not only stores the 3D geometry of the parts but also includes manufacturing-related information such as machining features, tolerance allocation, and process datums. This model is a key bridge connecting the design and process ends. The process system automatically generates process plans by parsing the machining features in the physical model, and the design end automatically corrects the geometric design by receiving process backflow constraints from the physical model. The four-layer models are closely linked through a four-dimensional bidirectional traceability mapping relationship. Any data point in any model can be traced upwards to the original requirement constraints and downwards to the physical geometric features. Any data change at any level is automatically synchronized with the other three-layer models, and change records are permanently archived for traceability. Step S1 is the data foundation and logical cornerstone of the entire forward design methodology. In traditional solutions, requirements, functions, performance, and geometric data belong to different systems and are isolated from each other. The underlying data model stores the four types of data in the same data architecture, establishes a strong correlation, and achieves full-domain synchronization for changes in one place, improving data consistency by more than 90%. The four-dimensional bidirectional traceability mapping relationship allows any geometric feature of the final physical model to be traced back to the initial design requirements and functional requirements. When a failure occurs during commissioning, the source of the failure can be quickly located at which level, which link, and which original constraint, improving the efficiency of fault tracing by more than 80%. The hierarchical structure of the four-layer model allows performance deviations to be accurately located at a specific level (top-level requirements / component design / part manufacturing), providing data support for layered deviation-driven iteration.
[0024] In some preferred embodiments, step S2 includes: S21. Receive external user requirements for the entire machine, call the design requirement model and functional model to complete the three-level layer-by-layer decomposition of system, component, and part, transform qualitative requirements into calculable quantitative constraints, and output the initial overall performance scheme of the entire machine; Specifically, external user requirements (such as "engine thrust not less than XX kN under altitude XX meters and Mach number XX" and "full life cycle fuel consumption not exceeding XX tons") are entered into the design requirement model after being structured. The system decomposes the top-level requirements into system-level constraints (such as overall machine pressure ratio and turbine inlet temperature), component-level constraints (such as compressor stage pressure ratio and turbine stage expansion ratio), and part-level constraints (such as blade size tolerance and material performance requirements) according to the preset requirement decomposition rules. S22. A four-dimensional traceability mapping relationship is established based on a four-layer model, so that changes in top-level requirements are automatically propagated down to the component and part levels. Specifically, when top-level requirements change (such as an upward adjustment of thrust indicators), the design requirement model automatically triggers the change propagation mechanism: first, the system-level quantitative constraints are updated; then, the relevant functional indicators are updated through the functional model; then, the relevant simulation conditions are recalculated through the performance model; and finally, the relevant geometric and tolerance requirements are updated through the physical model. The entire propagation process is completed automatically, and change records are retained throughout the entire process. S23. All decomposed quantitative indicators and functional constraints are synchronously updated to the underlying model of step S1 through the digital mainline, retaining a complete record of requirement changes. It is understandable that step S2 is the key bridge connecting user needs and engineering implementation. Existing technical solutions rely on the personal experience of designers for requirement decomposition and lack standardized methods. This step ensures the completeness and consistency of requirement decomposition through structured requirement decomposition rules and automated transmission mechanisms, avoiding requirement omissions and misunderstandings. In existing technical solutions, after requirement changes, each discipline needs to manually readjust design parameters, which is inefficient and prone to errors. The automatic transmission mechanism enables top-level requirement changes to be transmitted throughout the entire chain within minutes, improving change traceability efficiency by more than 80%.
[0025] In some preferred embodiments, step S3 includes: S31. The hierarchical quantization constraints output from step S2 are read from the digital mainline. A distributed parallel simulation architecture is adopted, with each component simultaneously performing multidisciplinary coupled calculations including aerodynamic and thermodynamic simulations, structural strength verification, rotor vibration analysis, overall weight estimation, and development cost calculation. Specifically, the compressor component team, combustion chamber component team, turbine component team, structural strength team, and rotor dynamics team conduct their respective multidisciplinary simulation calculations within the same time window, based on unified quantization constraints and boundary conditions. The distributed parallel simulation architecture achieves concurrent execution of multiple tasks through a high-performance computing cluster, significantly shortening the simulation cycle. S32. Based on a multi-objective optimization algorithm, find the Pareto optimal solution set that simultaneously satisfies all constraints of thrust, fuel consumption, lifespan, weight, and manufacturing cost. Specifically, the multi-objective optimization algorithm performs a global search in the parameter space of the performance model to find a solution set that achieves Pareto optimal balance among multiple objective functions (maximizing thrust, minimizing fuel consumption, maximizing lifespan, minimizing weight, and minimizing cost). Each solution in the Pareto optimal solution set represents a design scheme that cannot be surpassed by any other solution at all objectives. Researchers can select the most suitable scheme from these solutions based on project priorities. S33. Write back all simulation performance and functional parameters of the Pareto optimal solution set obtained in step S32 to the performance model and functional model of step S1 in both directions, and update the unified digital mainline data source; write back all simulation performance and functional parameters of the preferred scheme to the performance model and functional model of step S1 in both directions, and update the unified digital mainline data source. The distributed parallel simulation architecture in step S3 enables various components and disciplines to work simultaneously. Compared with the long serial process cycle of overall design → component design → verification → modification in existing technical solutions, the overall forward development cycle is shortened by more than 30%. Furthermore, the Pareto optimal solution set provides multiple equilibrium schemes, which R&D personnel can flexibly choose according to the actual needs of the project.
[0026] In some preferred embodiments, step S4 includes: S41. Read the optimal performance parameters output in step S3, and call the physical model template from step S1 to complete the 3D engineering modeling, dimensioning, and tolerance chain allocation of the parts; specifically, based on the optimal performance parameters, the system automatically calls the standard part template and typical structure template in the physical model template library to quickly generate the 3D geometric model of the parts, and automatically completes the dimensioning and tolerance chain allocation. The tolerance chain allocation is automatically optimized based on the accuracy requirements in the design requirement model and the manufacturing process capability database. S42. Receive the process quantification constraints returned from step S5 in real time, and automatically correct the model geometry, machining tolerances, and assembly datum. It can be understood that when the process feasibility verification in step S5 finds manufacturing problems, the system automatically generates quantification optimization constraints, which are returned to step S4 through the digital mainline to drive the physical model to be automatically corrected. S43. The corrected complete physical model data is synchronized back to the data model in step S1 to complete the version archiving.
[0027] In some preferred embodiments, step S5 includes: S51. Identify all machining geometric features within the physical model of step S4, and generate standardized machining process routes, tooling schemes, tolerance allocation strategies, and trial production cost calculation results in batches; specifically, the system automatically parses the machining features (such as surfaces of revolution, planes, hole systems, grooves, threads, tooth profiles, etc.) in the physical model through feature recognition algorithms, and automatically generates the optimal machining process route based on the pre-stored process knowledge base and tooling database. S52. Verify whether the process feasibility, processing cost, and tolerance processing accuracy meet the preset thresholds; if they do not meet the preset thresholds, automatically generate quantitative geometric modifications, tolerance adjustments, and structural simplification optimization constraints, and return to step S4 through the digital mainline to drive the iterative correction of the physical model, forming an automated design-process pre-process small closed loop, and cyclically verifying until the standards are met; if the preset thresholds are met, the complete process plan and cost assessment data are synchronously written back to the physical model in step S1, and the version record is permanently archived.
[0028] S53. All iterative modifications, version changes, and verification logs are uniformly stored in the underlying data model, supporting traceability and querying at any node throughout the entire process.
[0029] In some preferred embodiments, the design method further includes step S6: The test data is written back and compared with the simulation results in the performance model. The test data is collected and automatically compared with the simulation results in the performance model. Deviation optimization constraints are generated and pushed back to the design end to improve the global closed loop of the whole link. Specifically, the measured data of bench test (temperature, pressure, vibration, etc.) is automatically connected to the test data writing back module through the data interface. The module automatically compares the measured data with the simulation prediction data in the performance model, calculates the deviation amount and deviation distribution, and automatically traces the possible sources of deviation along the four-dimensional bidirectional mapping relationship. The system automatically generates optimization constraints and pushes them to the relevant models through the digital main line to drive the design end to make targeted corrections.
[0030] On the other hand, a preferred embodiment of the present invention also provides an end-to-end forward design system for aero-engines, comprising: The four-layer unified data model management module is used to create, associate, read, write, and version control the underlying data model. The requirement layering decomposition and traceability module is used to read the design requirement model to complete the three-level requirement decomposition, build a four-dimensional bidirectional traceability link, and automatically handle the global transmission of requirement changes. The parallel multidisciplinary evaluation and multi-objective optimization module is used to drive distributed multi-field coupled simulation and solve for Pareto optimal solution sets. The engineering modeling and geometry correction module is used to complete 3D modeling based on performance parameters and automatically correct geometry and tolerances upon receiving process reflow constraints. The design-process bidirectional closed-loop verification module is used to automatically analyze geometric machining features and generate process plans, and output manufacturing optimization constraints in reverse. The test data reverse write-back module is used to access data collected from bench tests and component tests, automatically benchmark against the simulation performance model, generate deviation optimization constraints, and push them back to the design end. The full-process version traceability and retrieval module is used to uniformly archive model changes, simulation records, process iterations, and test fault logs throughout the entire process. It is also used to support tracing back from test failures to the top-level original requirement constraints.
[0031] This system incorporates a fixed hierarchical mapping link: design requirement model (top-level requirements) → functional model (subsystem) → performance model (multidisciplinary components) → physical model (part geometry / tolerances). Any change in performance parameters, geometric dimensions, or requirement indicators will be tagged with a unique hierarchical label. Data boundaries at each level are completely isolated and can be accessed independently. The system automatically calibrates and verifies, tracing the source of deviations to determine their level. Performance deviation data from simulation, process, and testing is fed back into the four-layer model along a bidirectional digital thread. The system automatically matches the original constraint source based on the four-dimensional traceability mapping relationship, accurately determining that the deviation belongs only to a single level among the top-level requirement layer, component performance layer, and part physical geometry layer. The digital mainline shields irrelevant hierarchical model data, loading only the sub-model of the level to which the deviation belongs and the corresponding professional simulation solver. All other level models and whole-machine coupled calculation modules are paused and do not participate in the calculation. If the deviation only comes from the physical geometry layer of the part (such as blade machining tolerance, assembly clearance), only the part-level 3D geometry and local strength / aerodynamic simulation are called. There is no need to load the whole-machine multi-level flow channel, rotor dynamics, and combustion system simulation. If the deviation comes from the component performance layer (such as insufficient turbine aerodynamic efficiency), only the multi-field coupled simulation of the single component is carried out. The other components of the whole machine and the top-level overall scheme model are kept frozen and reused. Only when the top-level requirement index is set inappropriately will the full-level linkage iteration be triggered.
[0032] On the other hand, a preferred embodiment of the present invention also provides a dedicated electronic device for forward design of aero-engines, including a processor and a memory; the memory stores the underlying data model and the full-link forward design calculation program corresponding to any of the methods described in the computer execution program; the processor loads and executes the program to complete all the steps of any of the methods described.
[0033] On the other hand, a preferred embodiment of the present invention also provides a computer-readable storage medium storing a computer-executable program, which, when loaded and run by a processor, executes any of the aero-engine end-to-end design methods described above.
[0034] In the description of this invention, it should be noted that the terms "upper", "lower", "front", "rear", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.
[0035] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0036] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for end-to-end design of aero-engines, characterized in that, include: S1. The integrated underlying data model of components includes related design requirement models, functional models, performance models and physical models. All data in the model is synchronized bidirectionally to a unified digital mainline in real time. S2. Decomposition of requirements and bidirectional model linkage; Receive external whole machine user requirements, call the design requirement model and functional model to complete the three-level layer-by-layer decomposition of system, component and part, establish a four-dimensional traceability mapping relationship, and synchronously update to the underlying data model of step S1. S3. Parallel multidisciplinary collaborative design of components and optimal solution set solving; including distributed parallel simulation, Pareto optimal solution set solving and bidirectional write-back; S4. Component engineering drawings and physical geometry modeling correction; S5. Automatic mapping and verification of the design process: identify all machining geometric features in the physical model established in step S4 and generate process routes. Automatically verify whether the feasibility of the process meets the preset threshold. If the threshold is not met, automatically generate quantitative optimization constraints and backflow them to step S4 through the digital main line.
2. The end-to-end design method for aero-engines according to claim 1, characterized in that, The design requirements model stores the qualitative and quantitative requirements of the entire machine in a structured manner, including three-level quantitative constraints at the system, component, and part levels, rules for the transmission of requirement changes, and original logs for full-link traceability. The qualitative and quantitative requirements include top-level thrust, fuel consumption, lifespan, weight, and cost. The functional model carries the functional definitions of the subsystems of the entire machine, functional logic mapping, and functional index decomposition relationships, supporting the automatic transmission of requirement changes to each module. The performance model integrates a standardized parameter library of multiple disciplines such as aerodynamics and thermodynamics, combustion, structural strength, rotor vibration, and control, stores the simulation input boundaries, simulation output performance results, and performance verification thresholds at each level, supports distributed multi-field coupled joint simulation calculations, and stores the Pareto multi-objective optimization solution set. The physical model stores a complete set of geometric manufacturing information for parts, including 3D geometry, assembly constraints, machining features, tolerance chains, and process datum. It supports the automatic standardized mapping of the design model to the process system and synchronously receives geometric correction constraints from process backflow.
3. The end-to-end design method for aero-engines according to claim 1, characterized in that, Step S2 includes: S21. Receive external whole machine user requirements, call the design requirement model and functional model to complete the three-level layer-by-layer decomposition of system, component and part, transform qualitative requirements into calculable quantitative constraints, and output the initial whole machine overall performance scheme. S22. Establish a four-dimensional traceability mapping relationship based on the four-layer model, so that top-level requirement changes can be automatically transmitted down to the component and part levels; S23. All the decomposed quantitative indicators and functional constraints are synchronously updated to the underlying model of step S1 through the digital mainline, and a complete record of requirement changes is retained.
4. The end-to-end design method for aero-engines according to claim 3, characterized in that, Step S3 includes: S31. Read the hierarchical quantization constraints output from step S2 from the digital mainline, and adopt a distributed parallel simulation architecture to simultaneously carry out multi-disciplinary coupled calculations such as aerodynamic and thermodynamic simulation, structural strength verification, rotor vibration analysis, whole machine weight estimation, and development cost calculation for each component. S32. Solve the Pareto optimal solution set that simultaneously satisfies all constraints of thrust, fuel consumption, lifespan, weight, and manufacturing cost based on a multi-objective optimization algorithm; S33. Write back all simulation performance and function parameters of the Pareto optimal solution set obtained in step S32 to the performance model and function model in step S1 in both directions, and update the unified digital mainline data source.
5. The end-to-end design method for aero-engines according to claim 4, characterized in that, Step S4 includes: S41. Read the optimal performance parameters output in step S3, and call the physical model template in step S1 to complete the 3D engineering modeling, dimensioning, and tolerance chain allocation of the parts; S42. Receive the process quantification constraints from step S5 in real time and automatically correct the model geometry, machining tolerances, and assembly datum. S43. The corrected complete physical model data is synchronized back to the data model in step S1 to complete the version archiving.
6. The end-to-end design method for aero-engines according to claim 5, characterized in that, Step S5 includes: S51. Identify all machining geometric features within the physical model of step S4, and generate standardized machining process routes, tooling schemes, tolerance allocation strategies, and trial production cost calculation results in batches; S52. Verify whether the process feasibility, processing cost, and tolerance processing accuracy meet the preset thresholds; S53. All iterative modifications, version changes, and verification logs are uniformly stored in the underlying data model.
7. The end-to-end design method for aero-engines according to claim 1, characterized in that, The design method also includes step S6: Experimental data is written back and compared with the simulation results in the performance model; experimental data is collected, and the simulation results in the performance model are automatically compared with the simulation results in the performance model. Deviation optimization constraints are generated and pushed back to the design end.
8. A fully integrated design system for aero-engines, characterized in that, include: The four-layer unified data model management module is used to create, associate, read, write, and version control the underlying data model. The underlying data model includes the associated design requirement model, functional model, performance model, and physical model. The requirement layering decomposition and traceability module is used to read the design requirement model to complete the three-level requirement decomposition, build a four-dimensional bidirectional traceability link, and automatically handle the global transmission of requirement changes. The parallel multidisciplinary evaluation and multi-objective optimization module is used to drive distributed multi-field coupled simulation and solve for the Pareto optimal solution set; The engineering modeling and geometry correction module is used to complete 3D modeling based on performance parameters and automatically correct geometry and tolerances by receiving process reflow constraints. The design-process bidirectional closed-loop verification module is used to automatically analyze geometric machining features and generate process plans, and output manufacturing optimization constraints in reverse. The test data reverse write-back module is used to access data collected from bench tests and component tests, automatically benchmark against the simulation performance model, generate deviation optimization constraints, and push them back to the design end. The full-process version traceability and retrieval module is used to uniformly archive model changes, simulation records, process iterations, and test fault logs throughout the entire process. It is also used to support tracing back from test failures to the top-level original requirement constraints.
9. A dedicated electronic device for forward design of an aero-engine, characterized in that, It includes a processor and a memory; the memory stores a computer-executable program and a low-level data model and a full-link forward design operation program corresponding to the method of any one of claims 1 to 7; the processor loads and executes the program to complete all the steps of the method of any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The system contains a computer-executable program, which, when loaded and run by a processor, executes the end-to-end design method for aero-engines as described in any one of claims 1 to 7.
