Modal analysis and structural optimization of integrated bladed disks and centrifugal compressor impellers

Abstract

Reducing carbon emissions is critical, and Carbon Capture, Storage, and Utilization (CCSU) technologies can play a vital role. However, the energy needed for compression is a major obstacle to the success of these technologies. Thus, designing compressors with optimal aerodynamic and structural efficiency is essential. Therefore, the present work proposes an application of a structural optimization process for centrifugal compressors to achieve a more efficient and failure-resistant machine. Our optimization process primarily focuses on minimizing resonance risks, ensuring the static structural integrity. We employ pre-stressed modal analysis, accounting for factors like inertial loads, centrifugal stiffening, spin-softening, and gyroscopic/Coriolis effects. Objective functions for optimization are based on Campbell Diagram. Gaussian Process Regression is employed, and the training process is conducted iteratively using an implemented adaptive Bayesian Sampling method. Three optimization algorithms are utilized: the Genetic Algorithm, Particle Swarm Optimizer, and the Grey Wolf Optimizer. To validate our approach, we conduct three case studies, including two CO2 centrifugal compressors intended for CCSU projects. In the axial blisk case, we successfully eliminated all resonance conditions, achieved a 23% reduction in mass, and maintained stress levels similar to the baseline. For the first-stage centrifugal compressor, we increased the bursting margin by 4.31% and eliminated critical resonance conditions. In the fourth-stage case, we improved low-frequency resonance conditions, although without a significant enhancement in structural integrity compared to the baseline design. Despite the reduction in resonance risk, frequency tuning for the centrifugal compressors proved to be a challenging task, and further improvements are still necessary.