In the present study, an inverse design method has been developed to optimize and improve the performance of centrifugal compressors based on the performance curve. The core idea of this method is to create an accurate surrogate model for predicting compressor performance and combine it with multi-objective optimization techniques to enhance performance across a range of operating points. In this approach, the surrogate model predicts the compressor efficiency and pressure ratio according to its geometric parameters with sufficient accuracy, and the genetic algorithm iteratively modifies the compressor geometry until the target performance curve is achieved. This approach allows the centrifugal compressor geometry to be inversely designed based on the desired performance curve. The proposed method is also generalizable to other geometries, and problem variables, including geometric parameters and objective functions, can be redefined according to design requirements. The use of multi-point optimization in this approach ensures that the designed compressors perform satisfactorily not only at the design point but also at off-design operating conditions.

The results of this study indicate that the compressor efficiency improved by up to 11% and the pressure ratio by up to 14% at certain operating points. Accordingly, the developed design algorithm provides a reliable and accurate means to enhance the performance of centrifugal compressors and can serve as a fast and trustworthy method for the inverse design of centrifugal compressors based on a desired performance curve.