A semi-empirical method for assessing the performance of an open-drive screw refrigeration compressor

Abstract

A semi-empirical modeling procedure for twin-screw compressor performance simulation is proposed in this paper. In detail, the compression process is split into a number of stages, taking into account fluid leakages, heat transfers as well as heat and power losses. Compared to the simplified models reported in technical literature for other positive-displacement units, a different formulation for ambient heat loss is proposed, while mechanical power losses are modeled according to torque losses depending on both load and viscous friction effects. The proposed model applies to open-drive compressors, so three specific units for refrigeration applications with medium-high evaporation temperature are selected as the test cases for model tuning and validation, based on manufacturer's catalog data. After the identification of eight parameters, the model is able to compute the mass flow rate, the shaft power and the fluid discharge temperature, as well as volumetric and compressor efficiencies and ambient heat losses.Contrary to geometric models, the semi-empirical procedure proposed in this study does not take the presence of lubricating oil into account, as (i) oil injection with no external cooling results in a thermal neutral process; (ii) the compression work for the oil is negligible in comparison with the one required by the refrigerant fluid; and (iii) even injecting oil at a lower temperature, no considerable saving in refrigerant specific work can be achieved. Actually, this hypothesis does not significantly affect the output of the model. As a matter of fact, referring to the first test case with the larger displacement volume and to the identification of the eight model parameters based on eight convenient data points, the mass flow rate delivered by the compressor is predicted with a maximum error up to 2.44%, and the shaft power is predicted with an even limited error, except for few points with the lowest and highest pressure ratios, where the shaft power is under-estimated by 5%. Moreover, fluid temperature at the compressor exhaust is calculated with an error within ±1 K for pressure ratios up to 7.5. Similarly interesting results are achieved for the other two test case compressors, with absolute values of the errors in mass flow rate and shaft power predictions always less than 3% and 5% respectively.Finally, further considerations on the actual possibility of generalizing the model to semi-hermetic units are suggested.