Efficient metamodeling and uncertainty propagation for rotor systems by sparse polynomial chaos expansion

The modeling of rotor systems involves various parameters prone to uncertainties. These variations typically arise from the mathematical complexities of representing rotor system peculiarities and the limited understanding of material properties in specific applications. Analyzing uncertainties affecting rotor system performance is essential for effective design. A metamodeling approach for rotor systems under uncertain parameters is developed, employing sparse polynomial chaos expansion (sPCE) for uncertainty propagation. The sPCE method integrates basis functions adaptively using the Bayesian compressive sensing (BCS) method, enhancing convergence speed for accurate prediction of statistical moments. Probabilistic outcomes are compared with traditional Monte Carlo simulation (MCS) and Latin Hyper Sampling (LHS) methods. The comparative analysis shows that the proposed method achieves higher computational accuracy than the LHS method and exhibits a 40% improvement in computational efficiency compared to the traditional MCS method, thus providing valuable insights for the design and maintenance of rotor systems.