Compressor Cascade Optimization, BBD

Project Overview

A high-speed compressor cascade wind tunnel is utilized to investigate secondary flow phenomena in the corner and sidewall regions of axial compressors.

This project focuses on optimizing a compressor cascade using the Multi-Objective Genetic Algorithm (MOGA) method. Initially, we simulated a sectional compressor cascade configuration.

Subsequently, we performed optimization involving three input parameters and two output parameters. The input parameters include inlet velocity (v_in), angle of attack (alpha_degree), and pitch. The output parameters are drag force and lift force.

The objective function is defined to minimize drag force toward zero, maximize lift force to 0.07, and achieve a beta angle of -12 degrees.

Geometry and Meshing

The geometry was created as a 3D model using DesignModeler software. A computational grid was generated using ANSYS Meshing software, featuring an unstructured mesh with tetrahedral cells and 5 boundary layers. The total mesh contains 991,872 cells.

Optimization Methodology

All optimization procedures were executed in ANSYS Workbench software using Multi-Objective Genetic Algorithm (MOGA).

The Box-Behnken Design (BBD) method was implemented for the Design of Experiments (DOE) stage, while Genetic Aggregation served as the Response Surface Method (RSM). The input parameter ranges are defined as follows:

• Inlet velocity: 3 to 30 m/s
• Pitch: 1 to 7 mm
• Angle of Attack: -10° to +10°

Results and Analysis

Results were obtained at each of the three main optimization stages for analysis and optimal point selection. A summary table presents the design points and their corresponding execution results.

ANSYS Workbench utilized the DOE results to generate response surfaces for predictive analysis. Sensitivity charts illustrate how output parameters respond to variations in input parameters.

The analysis reveals that all three input parameters positively influence lift force, with velocity demonstrating the strongest effect.

While velocity positively affects both lift and drag forces, it negatively influences the beta angle. Conversely, pitch exhibits the most significant impact on the beta angle.

Optimization Outcomes

Upon completion of the optimization process, ANSYS Workbench identified three candidate points as optimal solutions, along with three verification points for validation.