Large size reciprocating compressor analysis with a Finite Volume 1D model

In the reciprocating compressor design process, the performance predictability plays a key role since the early stages. Performance of reciprocating compressors are often calculated by means of lumped parameter numeri-cal model, neglecting the presence of unsteady phenomena such as pressure waves travelling inside cylinder bore and suction/discharge gas chambers. Dealing with a large size compressor, it can be observed that the time needed for the pressure waves to travel the whole cylinder bore is comparable with the duration of the physical phenomena of the thermodynamic cycle. This affects the gas transfer process inside the cylinder chamber, mostly during the suction and discharge phases, and thus the compressor performance. The authors inferred that a spatial discretization of the compression chamber domain could be useful in order to more accurately predict the indicated power and mass flow rate. In this case, the accuracy of lumped parameter models in pre-dicting the compressor performance demonstrated to be improvable. In fact, the modelling of the compression chamber domain by using one only lumped element does not allow to simulate the gas motion inside the cylin-der. Therefore, a more realistic way to simulate the mass flow inside the chamber is to increase the spatial di-mension of the fluid domain representation. This leads to a numerical model that includes a spatial definition of the cylinder chamber, in order to account for the fluid inertial effects. In order to meet this need and to increase the accuracy of the thermodynamic cycle simulation, the authors developed a one-dimensional compressor model that solves the fluid dynamic equations by applying the finite volume method (FVM). At first, the FVM model was validated on a simplified test case simulated with transient CFD. Then, a real large size reciprocating compressor was tested both with the FVM model and a lumped pa-rameter model. The compressor performance was monitored and the numerical results were compared with experimental measurements collected on a dedicated test bench. The FVM model showed a better agreement with the experimental data with respect to the results of the lumped parameter model. The comparison between the numerical results and the experimental data was useful to assess the influence of the fluid distribution in the cylinder chamber on the compressor performance predictability. The importance of simulating the cylinder chamber with a good spatial discretization showed to be crucial to match the experimental data.