Multidisciplinary Design of a Low-Noise Propeller: Part II — Efficient Aero-Acoustic-Mechanical Design Methodology Exploiting Surrogate Models in an Adaptive Design Space

The regional aircraft segment plays a crucial role in achieving the EU Flightpath 2050 objectives (increase connectivity through Europe, enforce Europe's industrial leadership, and significantly reduce the environmental impact of aviation). Despite an outdated perception by the general public, turboprop aircraft are typically less expensive to operate than regional jets. The impact of new technologies is therefore even more evident. Achieving a significant reduction in perceived noise levels remains however a challenge for the success of further turboprop deployment. This twofold paper discusses the design of an innovative low-noise propeller in the framework of the Clean Sky 2 Regional Aircraft IADP, with a focus on the design methodology itself in this second part. The design is inherently multidisciplinary - aerodynamic, acoustic, mechanical - with multiple flight conditions and a wind tunnel condition to be considered. In order to limit the number of expensive high-fidelity computations, an online surrogate-based optimisation (SBO) approach has been deployed. A high-dimensional design space has been considered to enable to identify disruptive low-noise concepts. By exploiting the results of low-fidelity tools (see the first part of the paper), combined with efficient machine learning techniques and data mining capabilities, a gradual increment of the design space from 57 to 111 design parameters has been considered. A significant noise reduction of about 6.5 dB has been achieved without major degradation of the aerodynamic efficiency - fully aligned with the objectives for the Regional Aircraft IADP.