Application of New Empirical Models based on Mathematical Statistics in the Through-Flow Analysis

The through-flow analysis is an important part in the compressor design process. In this part, lots of empirical models are used to predict the compressor aerodynamic performances. These models are all based on the early experimental results of low-load and low-speed compressors; so, they may not be suitable for modern compressor performance estimations. In this study, several empirical models were reviewed, and two new models to estimate the minimum blade profile loss and minimum-loss deviation angle were established using mathematical statistics. Then, the original models with poor accuracy were replaced by these two models in through-flow program. Next, three compressors (Rotor 67, Stage 35, and a 4-stage repeating stage) were simulated using both the original and new models. The final results show that the new models have better accuracy in estimating the performance parameters than the original models. This indicates that it is feasible to use mathematical statistical methods to establish the empirical models, and this method provides an idea for the improvement of other empirical models in through-flow program.