Aircraft Propulsion Structural Design From Historical Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing aircraft propulsion system design processes lack efficiency in accommodating design constraints and often rely on conventional templates that may not be optimal for new designs, limiting innovation opportunities.

Innovation Solution

Utilizing an artificial intelligence model trained on historical geometry and operational data to generate a preliminary design of the propulsion system structural architecture, incorporating geometric and operational parameters, and performing finite element method analysis to ensure compliance with technical and customer constraints.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional template-based design methods are used, then design process simplicity is maintained, but design efficiency and innovation capability deteriorate

Engineering Contradiction:
Improvedesign efficiencyVSAvoiddesign process complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent replaces conventional template-based mechanical design processes with an AI-based generative design system. The AI model learns from historical design data and automatically generates optimized structural architectures, substituting manual iterative design with intelligent automated design that explores broader design spaces while satisfying constraints.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent transforms design from fixed template parameters to dynamically optimized parameters. The AI model identifies and optimizes geometric parameters based on learned patterns from historical data, allowing continuous parameter adjustment to meet specific constraints rather than relying on predetermined template values.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If AI-based generative design is implemented, then design optimization and constraint satisfaction improve, but computational complexity increases

Engineering Contradiction:
Improveconstraint satisfactionVSAvoidcomputational system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent performs preliminary training of the AI model on extensive historical design data before actual design generation. This pre-learning phase establishes the model's understanding of design constraints and optimal solutions, enabling it to efficiently generate reliable designs without requiring complex real-time computations during the actual design process.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses the AI model to generate design solutions by learning patterns from historical design data. Instead of relying on complex analytical computations for each design problem, the system copies successful design patterns and adaptations from historical data, adapting them to new constraint scenarios through the trained model.

Inventive Principle:
Principle #26Copying

3Manufacturing precision

If historical data training is performed, then design quality and best practice adherence improve, but data processing requirements increase

Engineering Contradiction:
Improvedesign qualityVSAvoiddata processing load
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent extracts key geometric and operational parameters from historical design data during the training phase. The AI model identifies and learns from essential design features and constraint-satisfaction patterns, extracting only the critical information needed for generating high-quality designs rather than processing entire historical datasets during actual design operations.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP4617936A1System and method for generating a preliminary design of a structural architecture for an aircraft propulsion system
Publication Date: 2025.09.17 PRATT & WHITNEY CANADA CORP
  • EP4617936A1 patent drawingFigure 1
  • EP4617936A1 patent drawingFigure 2
  • EP4617936A1 patent drawingFigure 3A

AI summary

A method for generating a preliminary design of a propulsion system structural architecture for an aircraft propulsion system includes generating, with an artificial intelligence (AI) model (600) at a computer system (400), one or both of at least one two-dimensional image and/or at least one three-dimensional model of the preliminary design including a selected combination of geometric parameters (804). The AI model has been trained for both a propulsion system geometry using the geometric parameters extracted from historical geometry data (408A) and a propulsion system structural design using the geometric parameters and operational parameters of historical operational data (408B). The operational parameters are associated with the geometric parameters. Training the AI model included identifying the selected combination of geometric parameters of the preliminary design, with the AI model, by determining a plurality of combinations of the extracted geometric parameters and the associated operational parameters which satisfy each of at least one technical constraint (520B) and at least one customer constraint (520A) and selecting the selected combination of geometric parameters from the plurality of combinations of the extracted geometric parameters using the associated operational parameters.