Computational Fluid Dynamics (CFD) is a common and validated simulation method in research and industry for the analysis of fluid systems. In the past years, it has proven to become more and more applicable for modeling the flow physics inside positive displacement (PD) machines. The working chamber and thus the discretized flow domain of PD machines are changing in time, characterized by complex thermodynamics. Compressible fluids, real-gas properties and leakage flows with trans- or supersonic characteristics are phenomena which have to be accounted for in order to properly model the behavior of the machine. As CFD methods evolve in general, but also for the application of PD machines in particular, the numerical model can replace a prototype during early stages of the product development. The desired simulation approach should be able to deliver sufficient accuracy at a feasible effort in terms of computational time and manpower to create the numerical model.This paper presents the methodology of creating the numerical model for a sample screw compressor provided by Sullair for research purposes. It is a dry running two stage twin screw compressor running with air at a rated power range between 160 and 250 kW. The two stages are gear driven by the main shaft at rotational speeds between 1180 and 2100 rev/min. Each stage features different rotor profiles, where the first stage has a 4-6, the second stage a 5-7 lobe combination. The total pressure ratio of the two stages combined is up to 10:1. To enhance the performance of the compressor, discharged air from the first stage is cooled down before entering the second stage. A specific meshing method is used to model the size-changing working chambers between rotors and casing, where only hexahedral cells are used and mesh topology is constant. The model accounts for radial and axial clearances between rotors and stator, where rotors and stator are connected with interfaces. The transient simulation results are compared to experimental measurements for torque, and flow rate. Also discharge pressure and temperature after first and second stage are compared to the experimental results. In addition, the possibilities of the simulation are exemplified by the gathering of time- and space-resolved monitor points like temperature or pressure at distinct points within the compressor. Apart from direct comparison to the experiment, also a sensitivity study regarding the change of housing clearances is presented, as leakage flow has severe impact on the compressor performance. These clearances and the resulting leakages are often not exactly known whereas they also vary because of manufacturing tolerances or deformations due to the load on rotors and stator. Here, the numerical simulation can serve as a helpful tool to estimate the sensitivity and change of machine characteristics, which is hard to determine in the scope of experiments.
Read full abstract