Abstract:The intent of this work was to investigate the feasibility of developing machine learning models for calculating values of airplane configuration design variables when provided time-series, mission-informed performance data. Shallow artificial neural networks were developed, trained, and tested using data pertaining to the blended wing body (BWB) class of aerospace vehicles. Configuration design parameters were varied using a Latin-hypercube sampling scheme. These data were used by a parametric-based BWB configuration generator to create unique BWBs. Performance for each configuration was obtained via a performance estimation tool. Training and testing of neural networks was conducted using a K-fold cross-validation scheme. A random forest approach was used to determine the values of predicted configuration design variables when evaluating neural network accuracy across a blended wing body vehicle survey. The results demonstrated the viability of leveraging neural networks in mission-dependent, inverse design of blended wing bodies. In particular, feed-forward, shallow neural network architectures yielded significantly better predictive accuracy than cascade-forward architectures. Furthermore, for both architectures, increasing the number of neurons in the hidden layer increased the prediction accuracy of configuration design variables by at least 80%.

What problem does this paper attempt to address?

The problem this paper attempts to address is the development of machine learning models to achieve inverse design of Blended Wing Body (BWB) aircraft based on task requirements. Specifically, the goal of the research is to calculate the values of aircraft configuration design parameters by providing task performance data related to time series. Traditional direct optimization methods usually require multiple iterations to approach the target performance, while inverse design methods can directly calculate suitable aircraft configurations from fixed task objectives. ### Main Issues 1. **Feasibility of Inverse Design**: The research aims to verify whether machine learning models (especially shallow artificial neural networks) can be used to achieve inverse design of BWB aircraft based on task requirements. 2. **Mapping of Multidimensional Inputs and Outputs**: Traditional response surface models perform poorly when dealing with multidimensional inputs and multiple outputs, while artificial neural networks (ANNs) have advantages in handling high-dimensional, nonlinear data relationships. Therefore, the study explores the application of ANNs in inverse design. 3. **Prediction Accuracy of the Model**: The research evaluates the prediction accuracy of different neural network architectures (such as feedforward neural networks and cascade feedforward neural networks) and explores the impact of increasing the number of neurons in hidden layers on prediction accuracy. ### Research Background - **Traditional Aircraft Design**: The traditional Tube and Wing (TAW) aircraft design method is mature, but the requirements for performance and environmental impact are becoming increasingly stringent, necessitating the exploration of new concepts. - **Blended Wing Body Aircraft**: BWB is a new type of aircraft design that differs from traditional TAW, requiring the tight integration of wings, fuselage, engines, and tail to improve performance. - **Advantages of Inverse Design**: Inverse design methods can fix task objectives to calculate design parameters that meet these objectives, avoiding the potential mismatches that may exist in traditional direct optimization methods. ### Research Methods - **Data Generation**: A large number of BWB configurations were generated using the Latin Hypercube Sampling (LHS) method, and performance data for each configuration was obtained through performance evaluation tools. - **Neural Network Training**: The K-fold Cross-Validation method was used to train and test the neural network models to ensure the generalization ability of the models. - **Random Forest Validation**: The Random Forest method was used to evaluate the prediction accuracy of the neural network models on unknown design points. ### Key Findings - **Effectiveness of Neural Networks**: The study shows that shallow feedforward neural networks significantly outperform cascade feedforward neural networks in prediction accuracy for inverse design. - **Effect of Increasing Neurons**: For both architectures, increasing the number of neurons in hidden layers can significantly improve the prediction accuracy of configuration design parameters, by at least 80%. In summary, this paper successfully demonstrates the feasibility of inverse design of BWB aircraft based on task requirements by developing and validating machine learning models. This method not only improves design efficiency but also provides new ideas for future new aircraft designs.

Mission-Driven Inverse Design of Blended Wing Body Aircraft with Machine Learning

Estimation of Hypersonic Vehicle Weight Using Physics-Informed Neural Network Supported by Knowledge Based Engineering

Assessment on Critical Technologies for Conceptual Design of Blended-Wing-body Civil Aircraft

Enhanced data efficiency using deep neural networks and Gaussian processes for aerodynamic design optimization

An adaptive machine learning based optimization methodology in the aerodynamic analysis of a finite wing under various cruise conditions

Neural network design for data-driven prediction of target geometry for an aerodynamic inverse design algorithm

Support Vector Regression-based Multidisciplinary Design Optimization in Aircraft Conceptual Design

Simultaneous surrogate modeling and dimension reduction using unsupervised learning. Application to parametric wing shape optimization

Artificial neural network based inverse design: Airfoils and wings

An Improved Data-Driven Modeling Method for Aircraft Based on Prediction and Optimization

Optimization Design of Structural Layout Based on Techniques of Artificial Neural Network

Physics-Infused Machine Learning Based Prediction of VTOL Aerodynamics with Sparse Datasets

Database Self-Expansion Based on Artificial Neural Network: an Approach in Aircraft Design

Inverse Design of Airfoils Using Convolutional Neural Network and Deep Neural Network

Design Methods and Strategies for Forward and Inverse Problems of Turbine Blades Based on Machine Learning

Prediction and optimization of airfoil aerodynamic performance using deep neural network coupled Bayesian method

Design Optimization for the Entire Aircraft Structure of Civil Aircraft with Blended-Wing-body Layout

Multi-fidelity deep neural network surrogate model for aerodynamic shape optimization

Utilizing Artificial Neural Networks for Entry Vehicle Aerodynamic Characterization

A novel approach to inverse design of wind turbine airfoils using tandem neural networks

Rotor37 Aerodynamic Optimization: A Machine Learning Approach