Search

There are some inherent limitations to the performance of support vector regression (SVR), such as (i) the loss function, penalty parameter, and kernel function parameter usually cannot be determined accurately; (ii) the training data sometimes cannot be fully utilized; and (iii) the local accuracy in the vicinity of training points still need to be improved. To further enhance the performance of SVR, this paper proposes a novel model modification method for SVR with the help of radial basis functions. The core idea of the method is to start with an initial SVR and modify it in a subsequent stage to extract as much information as possible from the existing training data; the second stage does not require new points. Four types of modified support vector regression (MSVR), including MSVg, MSVm, MSVi, and MSVc, are constructed by using four different forms of basis functions. This paper evaluates the performances of SVR, MSVg, MSVm, MSVi, and MSVc by using six popular benchmark problems and a practical engineering problem, which is designing a typical turbine disk for an air-breathing engine. The results show that all the four types of MSVR perform better than SVR. Notably, MSVc has the best performance. [ABSTRACT FROM AUTHOR]

This paper investigates an optimization for the axisymmetric hub-endwall profile of a transonic fan designed for a civil high bypass ratio turbofan engine. A time-saving integrated design optimization method is proposed based on a new hub-endwall profile modification method in combination with a novel two-stage polynomial response surface. The profile modification method is proposed to change the profile of the baseline hub curve by moving three control points, so as to redesign the hub-endwall. The two-stage polynomial response surface is proposed to obtain an accurate high-order polynomial while requiring fewer sampling points, so as to reduce the computational cost. The results show that the integrated design optimization method is suitable and valuable. After optimization, the corner separation of the studied fan is weakened, and the stall margin of the core fan is also enlarged. Moreover, the aerodynamic performance of the core fan at the design speed is improved. [ABSTRACT FROM AUTHOR]

While having been proven its significant advantage in losing weight, twin-web turbine disc (TWD) is known to be the novel stage of development in the area of high-pressure turbine discs. There is a significant axial temperature gradient which should be taken into consideration for the asymmetric web structure to obtain the optimal shape of TWD. An expensive three-dimensional fluid-thermal analysis should be introduced and executed to obtain accurate temperature distribution. A kind of multidisciplinary and multifidelity optimization procedure is proposed for the simultaneous combining of two obvious advantages with the low cost of axisymmetric strength analysis and the high fidelity of three-dimensional fluid-thermal analysis and strength analysis. It is executed under the support techniques, including zero circumferential elasticity modulus method at the non-axisymmetric region during axisymmetric strength analysis and inverse distance weighted method with traversing searching radius. A supporting ring with double sides locating mode at the disc hub is designed to avoid two webs being close to each other and provides the inlet for the cooling gas into the inner cavity. The optimization result demonstrates that the proposed method is capable and has an efficient benefit due to the low-cost computing procedure for obtaining the optimal shape of TWD with asymmetric temperature distribution. [ABSTRACT FROM AUTHOR]

Due to the good performance in terms of accuracy, sparsity and flexibility, support vector regression (SVR) has become one of the most popular surrogate models and has been widely researched and applied in various fields. However, SVR only depends on a subset of the training data, because the 휖-insensitive loss function ignores any training data that is within the threshold 휖. Therefore, some extra information may be extracted from these training data to improve the accuracy of SVR. By using the least squares method, a new improved SVR (ISVR) is developed in this paper, which combines the characteristics of SVR and traditional regression methods. ISVR is based on a two-stage procedure. The principle of ISVR is to treat the response of SVR obtained in the first stage as feedback, and then add some highly nonlinear ingredients and extra linear ingredients accordingly in the second stage by utilizing a correction function. Particularly, three types of ISVR are constructed by selecting different correction functions. Additionally, the performance of ISVR is investigated through eight mathematical problems of varying dimensions and one structural mechanics problem. The results show that ISVR has some advantages in accuracy when compared with SVR, even though the number of training points varies. [ABSTRACT FROM AUTHOR]

Multidisciplinary Design Optimization (MDO) method has been widely investigated and applied in the aircraft engine design. In the conceptual and preliminary design phase of the aircraft engine turbine MDO, automatic modeling and mesh generation of the blend between the blade and the shroud are two key technologies. A robust algorithm should be able to perform well without the target recognition of the geometric feature of the blend. A strategy of hexahedral mesh regeneration based on the Topological Homeomorphism is presented in this paper, where improved parametric modeling can be obtained for the turbine blade and disk. The optimal solutions of low pressure turbine of an aircraft engine could be obtained to examine the feasibility and availability of automatic modeling and mesh regeneration. The simulation results demonstrate that the proposed method is capable of satisfying the requirement on material strength by optimizing the blade blend radius, where the total weight of blade and disk can be reduced by 5.882%. [ABSTRACT FROM AUTHOR]