Aero-engine development is a highly complex system engineering, which involves multiple coupled disciplines and components. With increasing demands in terms of performance, costs and environmental friendliness, it’s necessary for the commercial aero-engine to have an integrated design system, which can take the coupling effects of different disciplines into account. This paper focus on a multi-level multidisciplinary optimization framework presented for an aero-engine which include multiple objective functions. To demonstrate the framework, we present an optimization for a turbine fan aeroengine considering three complex coupled subsystems, six coupled disciplines, fifty four design variables, and nine constraints. Within the optimization, an overall performance subsystem, a one-stage fan subsystem, and a six-stage low pressure turbine subsystem are analyzed in detail.

The overall performance subsystem involves analysis of thermodynamic cycle, as well as assessment of weight, noise and oxynitride emissions. To achieve an acceptable compromise among these coupled disciplines and conflicting objectives, eight related overall performance subsystem parameters are chosen as design variables. In addition, temperature limits of materials and pressure ratio limits of engine components are considered and treated as constraints. The fan subsystem and the low pressure turbine subsystem both involve analysis of aerodynamic performance and structural mechanics. To achieve an optimal design with respect to the aerodynamic performance and structural mechanics, thirty three fan subsystem parameters and thirteen low pressure turbine subsystem parameters are selected as design variables respectively. In addition, some requirements for aerodynamics and structure strength are considered and selected as constraints.

The basic principle and characteristics of the collaborative optimization (CO) strategy are introduced in this paper. Besides, the CO strategy is employed to solve the large-scale optimization problem. Moreover, a newly improved collaborative optimization (ICO) strategy is also proposed in this paper. Furthermore, the performance of the CO strategy and that of the ICO strategy are compared in detail.

The radial based function (RBF) surrogate models have been incorporated in the optimization process to reconcile high-fidelity CFD and FEA structural simulations of the fan subsystem, which are time-consuming. For constructing the surrogate models, a sampling technique termed orthogonal array design is employed to generate 108 sample points.

The non-dominated sorting genetic algorithm (NSGA-II) is used to find the Pareto front solutions for the different objectives in the overall performance subsystem. The gradient-based sequential quadratic programming algorithm (NLPQL) is employed to find optimal solutions rapidly in the fan subsystem, as well as the low pressure turbine subsystem.

The results show that the CO strategy and ICO strategy both produce a potential improvements for multiple objectives, when compared with the initial scheme. Additionally, when comparing the ICO strategy with CO strategy, we find that the ICO strategy has better optimization results.

This content is only available via PDF.
You do not currently have access to this content.