In aircraft wings, aileron mass parameter presents a tremendous effect on the velocity and frequency of the flutter problem. For that purpose, we present the optimization of a composite design wing with an aileron, using machine-learning approach. Mass properties and its distribution have a great influence on the multi-variate optimization procedure, based on speed and frequency of flutter. First, flutter speed was obtained to estimate aileron impact. Additionally mass-equilibrated and other features were investigated. It can deduced that changing the position and mass properties of the aileron are tangible following the speed and frequency of the wing flutter. Based on the proposed optimization method, the best position of the aileron is determined for the composite wing to postpone flutter instability and decrease the existed stress. The represented coupled aero-structural model is emerged from subsonic aerodynamics model, which has been developed using the panel method in multidimensional space. The structural modeling has been conducted by finite element method, using the p-k method. The fluid –structure equations are solved and the results are extracted.

Aeroelastic conditions are the main features to be considered in an the design of Arial vehicle. In other words, due to the structural interaction of the Arial components with the aerodynamics, it potentially yields to a coupling between fluid and structures. Among the aeroelastic phenomena, self-vibrations, at a certain speed, which is called fluttering, are considered as the most common problem in coupled systems. This destructive phenomenon will lead to the aggravation of high-amplitude vibrations effect [

Several attempts to represent analyzing tools regarding the flutter behavior, in a different regime based on the computational theories, were performed [

Design procedures of airplane structures are completely influenced by control surfaces such as aileron and flaps. This influence has direct effect on wing performances and therefore is considered as one of the challenging problems. The aeromechanical design instructions can be revealed from the study on positioning and instability of the ailerons [

The finite element method (FEM) as another useful tool was more likely employed, especially in designing procedures, for flutter boundary analysis, finite oscillation, and thermal problems [

Among the optimization methods [

In the present paper, the effect of the aileron's position, as well as mass inertial moment on the speed and frequency of the flutter, have been revealed for the composite wing. In addition, FEM is employed for modeling composite wing structure along with aerodynamic panel theory in the purpose of seeking components of the wing structure. The aero-elastic model was solved using the P-K method through the Nastran software. The novel optimization procedure is proposed and applied to find the best position of the aileron with the minimum state of the TSAI–WU stress via USAR [

For the sake of a better optimization of the design, several factors have been taken into consideration: dimensions and weight of the aircraft; type of maneuvers (permanent or sudden); weather conditions; magnitude of forces applied. In addition, the wing structure must be able to withstand all different conditions and keep the stability and control with a suitable reliability factor to satisfy and provide a safe and secure flight. In brief, the design of the structure of an aircraft, especially its wing with control surface like aileron, should be implemented in such a way that it should bypass the flutter criterion under the minimum possible weight and stress, following the standard air regulations.

Parameter | Carbon/epoxy |
---|---|

t (mm) | 0.28 |

Dens (kg/mm^{3}) |
1.42E-6 |

Ex (MPa) | 44250 0 |

Ey (MPa) | 44250 |

Vxy | 0.037 |

Gxy (MPa) | 5000 |

Fxt (Mpa) | 442 |

Fxc (Mpa) | 243 |

Fyt (Mpa) | 442 |

Fyc (Mpa) | 243 |

Fsxy (Mpa) | 45 |

The finite element model of aerodynamics and wing structure is presented in

The boundary conditions for solving this problem, considering that the wing is completely attached to the body, are completely fixed, and are numbered one and two from the beginning of the spar.

In the purpose of a better wing structure implementation, it is necessary that the loads are applied in different maneuver conditions. This stipulation is added in order to cope all different scenarios, encountered in a normal flight and to ensure that the aircraft structure, including its wing and its facility, is able to withstand the worst loading conditions. The aerodynamic loading group determines the worst loading conditions obtained from the above conditions in the worst maneuver conditions, by applying the reliability coefficient in the standards. Different structural members are designed and their strengths are determined in accordance with paragraph JAR25–301 [

Flutter analysis is conducted by aero elastic section of the Nastran software module. The data required for the flutter analysis is obtained by modal wing analysis. The flow regime for the above- mentioned was selected to cope the worst-case scenario as the unstable regime is considered and the Mach number equal to 0.6 was assigned. The air density was 1.225 kg/m^{3}. For estimating the speed range, 1 to 300 meters per second was considered.

Investigation of the aileron effect on the speed and frequency of the fluttering as the main parameter of design criteria is represented in this section. The results are compared to what extracted by represented model [

Parameter | With aileron | Without aileron | |
---|---|---|---|

Flutter Speed (m/s) | Present | 210 | 190 |

[ |
209.8 | 189.7 | |

Flutter Frequency (Hz) | Present | 17.4 | 16.45 |

[ |
17.38 | 16.44 |

For sea-level elevation, the velocity-frequency and velocity-damping diagrams are shown as a sample for the wing, despite aileron in the reference position in

Parameter | Original aileron | Aileron with concentrated mass | Error percentage % |
---|---|---|---|

Flutter Speed (m/s) | 190 | 190 | 0 |

Flutter Frequency (Hz) | 16.45 | 16.15 | 0.6 |

According to the Genetic Algorithm (GA) and Artificial Neural Network (ANN), a novel MDO method is adopted to propose the best position of the aileron on the composite wing, in order to postpone flutter and alleviated the stress of the root. The design flow includes parameterization aileron position, optimization algorithm, and a surrogate model on FEM (NASTRAN) software. Design of experiments (DOE) is also employed to create a database composed of the main mentioned composite wing with aileron along with Neural Network Algorithm. Following the created database, the flutter response and TSAI–WU stress criteria of the composite wing are evaluated. The new aileron positions are extracted using numerical calculation. The database is composed of the ANN results that are converged to numerical results. Finally, using the results of the NASTRAN software, the objective function is examined to assess the target goal satisfaction.

The minimum stress due to the above worst-case loadings along with flutter avoidance criteria makes the design optimization algorithm straightforward. Indeed, the procedure is followed to minimize the stress

Based on the meta-models idea besides ANN and also through the FEM aero-mechanical calculations of the original composite wing, DOE method is considered here. The network is trained via a feed forward-back propagation network with 8 hidden layers and one output neuron. Based on the experience, the proper range of the aileron position is set to be in a 5% deviation of the original position. ANN is adopted along with the approximated function

According to the DOE, the composite wing is followed by 50 Latin Hypercube types; 75% of the experiments are employed for network training and the other data are used for network validation. Here, the values of precision for efficiency are 99.9% with a 0.09% deviation. Based on FEM calculations, the results are compared with the approximation of the network in each ANN loop to assess the precision of the network in aileron position prediction. Using 35-generation and 70 members in each generation, the process of optimization is performed. If the precision of three-dimensional FEM simulations is less than 0.3%, the results of the neural network are applied for following the optimization procedure. The speed and frequency of the flutter of the best-predicted position of the aileron are represented in

The estimated stress results are compared with the main composite wing to reach the location of the aileron. The normalized TSAI–WU stress values for the all-represented position show normalized stress in the range of 0.45 to 1.2.

Multi-disciplinary optimization method represented number 5 (

Number | Aileron position relative to the middle axis of the body (meters) | Flutter speed (meters per second) | Flutter Frequency (Hertz) |
---|---|---|---|

1 | 1.95 | 215 | 16.7 |

2 | 2.66 | 195 | 16.2 |

3 | 3.38 | Flutter does not occur | - |

4 | 4.55 | Flutter does not occur | - |

5 | 5.94 | Flutter does not occur | - |

6 | 6.23 | 177 | 15.3 |

7 | 6.94 | 193 | 15.9 |

8 | 7.14 | 182 | 15.1 |

9 | 7.55 | 123 | 16.9 |

10 | 8.61 | Flutter does not occur | - |

11 | 9.22 | Flutter does not occur | - |

In this study, the effect of aileron position and its critical outcomes on the aeromechanics conditions of composite wing, with a high aspect ratio, were firstly investigated. Given that the speed and frequency of the flutter are directly related to mass distribution, the effects were so dramatic that in some situations the phenomenon of the flutter did not occur at speeds of up to 300 m/s. This study also showed that concentrated mass can be used instead of aileron's total modeling, and it was shown that there is no need to unify the inertial moment of mass inertia and, moreover, only the mass and positions of the center of mass are effective. Finally, based on the USAR [