Numerical investigation is conducted on a 3.5-stage axial compressor
on which numerous experimental projects were carried out at the Institute during the last years and an experimental database was established.In the current study five on-and off-design operating points are simulated using a RANS solver and the results are compared with the measurement.The result shows that the compressor performance can be qualitatively predicted by the mixing-plane method.Better agreement is obtained for the on-design operating point.However
as the flow unsteadiness is insufficiently considered
the numerical method produces end-wall low-speed flow layers accumulated with the flow passing through the passage
which is in no good agreement with the experimental data.In the numerical simulation the rotor rows receive less work and this difference from the measurement increases with the rotational speed.In contrast
the stator rows increase the pressure more efficiently than the measurement.In the simulation the flow in the last stator row tends more to separate on the pressure side of the blade.For the operating points close to the surge line
the predicted separation is more intense than the experimental observation.But for the operating points close to the choke
the separation is suppressed.
Abstract
Numerical investigation is conducted on a 3.5-stage axial compressor
on which numerous experimental projects were carried out at the Institute during the last years and an experimental database was established.In the current study five on-and off-design operating points are simulated using a RANS solver and the results are compared with the measurement.The result shows that the compressor performance can be qualitatively predicted by the mixing-plane method.Better agreement is obtained for the on-design operating point.However
as the flow unsteadiness is insufficiently considered
the numerical method produces end-wall low-speed flow layers accumulated with the flow passing through the passage
which is in no good agreement with the experimental data.In the numerical simulation the rotor rows receive less work and this difference from the measurement increases with the rotational speed.In contrast
the stator rows increase the pressure more efficiently than the measurement.In the simulation the flow in the last stator row tends more to separate on the pressure side of the blade.For the operating points close to the surge line
the predicted separation is more intense than the experimental observation.But for the operating points close to the choke