Parametric Rotor Control Equivalent Turbulence Input Models Using Neural Networks

Abstract

This paper extends the concept of Rotor Control Equivalent Turbulence Input (RCETI) to simulate the response of various rotor configurations to two-dimensional spatial turbulence, with the aim of developing parametric RCETI models. Using CIFER®, transfer functions were constructed to match the Power Spectral Density (PSD) of required inputs, effectively replicating the rotor response to turbulence. Unlike previous studies, this work encompasses a broader range of rotor parameters and employs non-dimensional parameters to study the generalizability of the model. Additionally, neural network modeling is utilized to create parametric models, with non-dimensional rotor parameters as inputs and RCETI model parameters as outputs. The study demonstrates that these parametric RCETI models can be effectively developed using a simple neural network architecture, which also shows scalability to different rotor configurations.