An Automated Geometric Analysis and Characterization of an Oil-Lubricated Twin-Screw Compressor for Predictive Modeling

Compressor manufacturing companies are showing a growing interest in the development of validated compressor simulation models capable of predicting the behaviour of the machine if subjected to appropriate modifications. A validated predictive model of the entire compressor is useful for analysing the influence of the variation of some fundamental parameters, such as the variation of volumes or new inlet and outlet sections, on the behaviour of the machine. To create these models, geometric parameters such as the variation in volumes, the variation in inlet and outlet sections and the presence of leakage or blowhole phenomena must be provided as input. The precision of these geometric parameters is of fundamental importance for the correct operation of the model and subsequent design analyses. In this work, the authors analysed the complete geometry of an oil-lubricated twin-screw compressor. In particular, an automated calculation methodology was developed using CAD software capable of defining the characteristic curves of the volumes and areas of the compressor: starting from the geometries of the rotors and the stator case, the fluid volumes delimited at the bottom by the rotors and at the top by the case were defined through a Boolean subtraction operation. Subsequently, the volumes that complete the compression phase together were isolated and, through an automatic kinematic analysis with steps of 5° of rotation, starting from the zero reference (null volumes), the characteristic curves of the volumes and inlet and outlet areas were defined through the use of special virtual sensors. The passage sections of the leakage and blowhole leaks were defined in the same way. Particular attention was paid to the analysis of leakage and blowhole phenomena, the presence of which determines a reverse flow of part of the mass flow rate of air during the compression phase, resulting in a decrease in volumetric efficiency. Knowledge of this data is of fundamental importance for the subsequent creation of the 1D/0D model.